fish behavior and abundance at the electric dispersal barrier in … · 2015-08-06 · fish...
TRANSCRIPT
Fish Behavior and Abundance at the Electric Dispersal Barrier in the Chicago Sanitary and Shipping Canal at Reduced and Current Voltage Operating Parameters
FEBRUARY 2015
Department of the Interior U.S. Fish and Wildlife Service Great Lakes-Big Rivers Region
Aaron D. Parker1,2, David C. Glover3, Samuel T. Finney1, P. Bradley Rogers1, Jeffrey G. Stew-art1, and Robert L. Simmonds Jr1
1U.S. Fish and Wildlife Service, Carterville Fish and Wildlife Conservation Office, 9053 Route 148, Marion, Illinois 62959 2Michigan Department of Environmental Quality, Surface Water Assessment Section, 525 W Allegan ST, Lansing, Michigan 48909 3The Ohio State University, Department of Evolution, Ecology, and Organismal Biology, 1314 Kinnear RD, Columbus, Ohio 43212
Disclaimer
The mention of trade names or commercial products in this report does not constitute endorse-ment or recommendation for use by the Federal government.
Nondiscrimination Clause
Equal opportunity to participate in and benefit from programs and activities of the U.S. Fish and Wildlife Service is available to all individuals regardless of race, religion, color, or physical or mental disability. For more information please contact: U.S. Department of the Interior Office of Equal Opportunity 1849 C. Street, N.W. Washington D.C. 20240
Mission Statement
The mission of the U.S. Fish and Wildlife Service is working with others to conserve, protect, and enhance fish, wildlife, plants and their habitats for the continuing benefit of the American people.
Suggested Citation
Parker, A. D., P. B. Rogers, J. G. Stewart, D. C. Glover, S. T. Finney, and R. L. Simmonds Jr. 2014. Fish Behavior at the Electric Dispersal Barrier in the Chicago Sanitary and Shipping Canal at Reduced and Current Voltage Operating Parameters. Region 3 Final Report 2015, U.S. Fish and Wildlife Service, Carterville Fish and Wildlife Conservation Office, Marion, Illinois.
II
Table of Contents Executive Summary ................................................................................................................................... - 1 -
Introduction .............................................................................................................................................. - 4 -
Materials and Methods ........................................................................................................................... - 13 -
Fish behavior and abundance near the barriers ............................................................................. - 13 -
Caged-fish trials ............................................................................................................................... - 18 -
Demonstration barrier caged-fish trials .......................................................................................... - 25 -
Statistical Analyses .................................................................................................................................. - 26 -
DIDSON sampling data analyses ..................................................................................................... - 26 -
Caged-fish data analysis .................................................................................................................. - 28 -
Results ..................................................................................................................................................... - 30 -
Former Operating Parameters (0.79 V/cm) Fish behavior and abundance near the barriers ....... - 30 -
Former Operating Parameters (0.79 V/cm) Caged-fish trials ......................................................... - 40 -
Current Operating Parameters (0.91 V/cm) Fish behavior and abundance near the barriers ...... - 45 -
Current Operating Parameters (0.91 V/cm) Caged-fish trials ........................................................ - 61 -
Discussion................................................................................................................................................ - 67 -
Acknowledgements ................................................................................................................................. - 79 -
References .............................................................................................................................................. - 80 -
III
Figures
Figure 1. Chicago Area Waterways System and location of electrical barriers. ....................................... - 5 -
Figure 2. Schematic of the electric barrier within the CSSC where DIDSON images were recorded. Note: drawing is not to scale and represents DIDSON recordings taken if Barrier IIB was operating. Sites were adjusted accordingly when Barrier IIA was in operation. ....................................................................... - 15 -
Figure 3. Cross-sectional illustration of the CSSC displaying ten sub-sites (A-J) across the canal where DIDSON footage was recorded within sites 1-8 from a north-oriented boat. Note: only one boat was used at a time to record one sub-site. .................................................................................................... - 16 -
Figure 4. Cage mounted on the side of an aluminum boat. ................................................................... - 20 -
Figure 5. Illustration depicting the cage and DIDSON mounting schemes on the north-oriented research boat in the CSSC and the two parts of the canal that the boat travelled. Note: only a single boat was used for caged-fish trials. ........................................................................................................................ - 20 -
Figure 6. Schematic of the electric barrier within the CSSC where caged-fish were moved through from downstream to upstream. Observation points used in the control runs were the same distances apart but outside of the barrier area. Observation points are where behaviors were later recorded. Note: drawing is not to scale and represents caged-fish observation points if only Barrier IIB was operating. Sites were adjusted accordingly if Barrier IIA was operating. ................................................................ - 23 -
Figure 7. Probability of individual fish presence at sites 1-8 within the electric dispersal barrier in summer 2011 (pooled across sections). Bars with the same letters indicate statistically similar odds ratio estimates between sites (P > 0.05). ........................................................................................................ - 31 -
Figure 8. Probability of schooled fish presence at sites 1-8 within the electric dispersal barrier in summer 2011 (pooled across sections). Bars with the same letters indicate statistically similar odds ratio estimates between sites (P > 0.05). ........................................................................................................ - 32 -
Figure 9. Probability of schooled fish presence at sections A-J within the electric dispersal barrier in summer 2011 (pooled across sites). Bars with the same letters indicate statistically similar odds ratio estimates between sections (P > 0.05). .................................................................................................. - 33 -
Figure 10. Probabilistic estimates of individual fish and schooled fish presence plotted across sites and sections within the electric dispersal barrier from summer 2011. ......................................................... - 34 -
Figure 11. Probability models for individual fish presence based on size at each site. Black lines represent predicted probabilities and gray lines represent 95% confidence intervals. ......................................... - 35 -
Figure 12. Probability models for schooled fish presence based on size at each site. Black lines represent predicted probabilities and gray lines represent 95% confidence intervals. ......................................... - 36 -
Figure 13. Percentages of different behaviors observed by individual fish at the different sites within the electric barrier in the CSSC...................................................................................................................... - 38 -
Figure 14. Percentages of different behaviors observed by schools of fish at the different sites within the electric barrier in the CSSC...................................................................................................................... - 39 -
IV
Figure 15. Predicted incapacitation probabilities of encaged gizzard shad as a function of distance at different locations with the barrier operating at 0.79 V/cm. Lines extending from x and y axes denote predicted point at which half of the fish became incapacitated (mid-channel and west wall only). Points represent the means (± SE) of observations at site distances, but the model is based on actual observations (0 or 1). Shaded areas represent locations of wide and narrow barrier arrays. ............... - 42 -
Figure 16. Percentages of different recorded behaviors for caged fish at each observation point in the control area. ............................................................................................................................................ - 43 -
Figure 17. Percentages of different recorded behaviors for caged fish at each observation point in the mid-channel of the electric barrier. ........................................................................................................ - 44 -
Figure 18. Percentages of different recorded behaviors for caged fish at each observation point along the west wall of the electric barrier. ....................................................................................................... - 45 -
Figure 19. Season-specific probabilities of individual fish presence at sites 1-8 within the electric dispersal barrier throughout 2012 (pooled across sections). Bars with similar letters indicate statistically similar odds ratio estimates between sites within a season (P > 0.05). An asterisks (*) indicates that individual fish were never observed at these sites such that no valid comparisons were possible. ..... - 47 -
Figure 20. Season-specific probabilities of schooled fish presence at sites 1-8 within the electric dispersal barrier throughout 2012 (pooled across sections). Bars with similar letters indicate statistically similar odds ratio estimates between sites within a season (P > 0.05). An asterisks (*) indicates that schooled fish were never observed at these sites such that no valid comparisons were possible. ...................... - 48 -
Figure 21. Season-specific probabilities of schooled fish presence at sections A-J within the electric dispersal barrier throughout 2012 (pooled across sites). Bars with similar letters indicate statistically similar odds ratio estimates between sections within a season (P > 0.05). An asterisks (*) indicates that schooled fish were never observed at these sections such that no valid comparisons were possible. . - 49 -
Figure 22. Season-specific probabilities of individual fish presence at sections A-J within the electric dispersal barrier throughout 2012 (pooled across sites). Bars with similar letters indicate statistically similar odds ratio estimates between sections within a season (P > 0.05). An asterisks (*) indicates that individual fish were never observed at these sections such that no valid comparisons were possible. - 50 -
Figure 23. Probabilistic estimates of individual fish presence plotted across sites and sections within the electric dispersal barrier for each season within 2012. The lack of replicated observation events in spring, summer, and fall 2012 resulted in site and section probabilistic estimates to be either 0 or 1.- 51 -
Figure 24. Probabilistic estimates of schooled fish presence plotted across sites and sections within the electric dispersal barrier for each season within 2012. The lack of replicated observation events in spring, summer, and fall 2012 resulted in site and section probabilistic estimates to be either 0 or 1.- 52 -
Figure 25. Probability models for fall season, individual fish presence based on size at each site. Black lines represent predicted probabilities and gray lines represent 95% confidence intervals. ................ - 54 -
Figure 26. Probability models for summer season, schooled fish presence based on size at each site. Black lines represent predicted probabilities and gray lines represent 95% confidence intervals. Note: no information is presented for site six because no fish schools were present at that time. ..................... - 55 -
V
Figure 27. Proportions of different behaviors observed by individual fish at the different sites within the electric barrier in the CSSC...................................................................................................................... - 56 -
Figure 28. Proportions of different behaviors observed by schooled fish at the different sites within the electric barrier in the CSSC...................................................................................................................... - 57 -
Figure 29. Mean sizes of schooled fish present at the zone of ultimate field strength (A) and mean dissolved oxygen concentrations (B) and temperatures (C). Each bar represents a single day of sampling in April, May, July, and September 2012. Hatch marks in dissolved oxygen and temperature bars indicate days in which a school of fish was present in the zone of ultimate field strength. ................................ - 59 -
Figure 30. Mean sizes of individual fish present at the zone of ultimate field strength (A) and mean dissolved oxygen concentrations. Each bar represents a single day of sampling in April, May, July, and September 2012. Hatch marks in dissolved oxygen bars indicate days in which individual of fish were present in the zone of ultimate field strength ........................................................................................ - 59 -
Figure 31. Proportions of different behaviors observed by individual and schooled fish during the spring and summer of 2012 at the zone of ultimate field strength in the barrier. ........................................... - 60 -
Figure 32. Predicted relationships of the incapacitation probabilities of encaged Gizzard Shad as a function of distance during different seasons and alongside different boat hulls (spring and summer only [during the fall and winter, only aluminum-hull boats were used]). Lines extending from x and y axes denote predicted point at which half of the fish became incapacitated. Points represent the means (± SE) of observations at site distances, but model is based on actual observations (0 or 1). ......................... - 63 -
Figure 33. Principal components analysis bi-plot of three continuous variables measured during trial runs. Arrows represent eigenvectors multiplied by two to scale to bi-plot area. .................................. - 64 -
Figure 34. Regressions of principal component 1 and barrier-run, incapacitation distances (natural-log transformed). Diamonds (♦) denote trial runs along the canal wall with an aluminum-hull boat, squares (■) denote trial runs along the canal wall with a fiberglass-hull boat, triangles (▲) denote trial runs through the mid-channel of the canal with an aluminum-hull boat, and circles (●) denote trial runs through the mid-channel of the canal with a fiberglass boat................................................................. - 65 -
VI
Tables
Table 1. Description of DIDSON recording sites. .................................................................................... - 16 -
Table 2. Description of caged-fish observation points. .......................................................................... - 24 -
Table 3. Summary of stepwise logistic regression models used to estimate the probability of individual and schooled fish presence/absence at the zone of ultimate field strength. ........................................ - 58 -
Table 4. Mean total length of gizzard shad used in trials, predicted distances at which 50% of the caged fish would be incapacitated, generalized coefficients of determination (R2) and estimated logistic regression slopes and intercepts under various conditions. Results are only for trial runs that took place at the current barrier operating parameters. ......................................................................................... - 62 -
Table 5. Correlation matrix of PC 1 and 2 scores with environmental and size variables. ..................... - 65 -
VII
Executive Summary
The Chicago Area Waterway System is an artificial network of heavily modified rivers
and canals that connect the Great Lakes and Mississippi River Basins. Because of concern about
inter-basin fish species exchange, an electric barrier is currently operated by the U.S. Army
Corps of Engineers in the Chicago Sanitary and Ship Canal 60 km south of Lake Michigan. The
electric barrier consists of two large barriers (only one of which was operated at any given time
during the study period) and a smaller Demonstration Barrier. The larger barrier operated at 6.5
ms, 15 Hz, 0.79 V/cm (0.79 V/cm) from 8/2009 to 11/2011 before being increased to 2.5 ms, 30
Hz, 0.91 V/cm (0.91 V/cm). These operating parameters were configured to inhibit the upstream
dispersal of Bighead Carp (Hypophthalmichthys nobilis) and Silver Carp (H. molitrix) from the
Mississippi River Basin to Lake Michigan. The U.S. Fish and Wildlife Service (Service) -
Carterville Fish and Wildlife Conservation Office began evaluating fish behavior at the electric
barrier in 6/2011 when the barrier was operating at 0.79 V/cm and continued to this work after
operating parameters were increased to 0.91 V/cm in 11/2011.
The Service’s efforts at the electric barrier consisted of a two-part study that involved ob-
servations of wild fish within the electric barrier and caged-fish behavioral trials at the barrier.
The abundances and behaviors of wild fish were evaluated near the electric barrier using a dual-
frequency identification SONAR (DIDSON) unit throughout the electric barrier. During the fall
of 2011, when the barrier was operating at 0.79 V/cm, fish accumulated below the electric barrier
and sometimes persistently challenged the electric field. Smaller fish were able to swim further
into the electrical field, than larger fish, to the area of water containing the ultimate field
strength. Caged-fish trials consisted of placing Gizzard Shad (Dorosoma cepedianum) in a non-
conductive cage that was attached to a boat and pulling the cage through the barrier. Eight of 270
fish that were pulled through the electric barrier did not become incapacitated.
After the barrier operating parameters were increased to 0.91 V/cm, we continued to
evaluate wild fish behavior within and around the barriers and continued our caged-fish work
throughout 2012. We found that wild fish at the electric barrier were scarce during the winter and
spring, and were not observed near the areas of highest voltage. However, wild fish accumulated
below the electric barrier the most during the summer and fall, and persistently challenged the
barrier. Fish that were able to penetrate the farthest into the electric barrier were smaller and
tended to aggregate at the water surface, near the canal walls.
Concentrated DIDSON sampling, in which all observations were made within the zone of
ultimate barrier field strength, revealed schools of wild fish persistently probing the barrier. Fish
were most abundant in this zone during the summer when both temperatures and dissolved oxy-
gen concentrations were high.
All caged-fish that were pulled through the barrier were incapacitated at some point with
the increased operating parameters of 0.91 V/cm. However, the distance that the fish were able to
swim varied considerably depending on season, size of fish, and the conductance of the boat that
they were moved next to (aluminum or fiberglass hull). Fish that were moved through the barrier
during the winter and spring seasons became incapacitated the soonest. The winter fish most
likely became incapacitated after a short distance because of reduced metabolic function in the
cold temperatures. The fish used in the spring were the largest fish used in the caged-fish trials
because only age-1, wild Gizzard Shad were available at the time, and were incapacitated at the
same distances into the barrier as the winter fish. When fish were moved through the barrier im-
mediately adjacent to the canal wall, incapacitations were delayed, indicating that the canal wall
- 2 -
may weaken or distort the electrical field to some degree. The fish that were moved through the
barrier in the summer showed the largest discrepancy in distances that they were able to swim
into the barrier, depending on the type of boat used to move them. Fish that were moved along an
aluminum-hull boat were able to swim nearly twice the distance into the barrier before becoming
incapacitated, as those that were moved along a fiberglass-hull boat. The delayed incapacitations
of fish moving along the aluminum-hull boat are presumably due to electrical distortion caused
by the metal boat conducting electricity. These results raise concerns about the effect that the
numerous metal-hull barges have on the barrier as they traverse it and the potential for them to
facilitate fish breach. The accumulation of fish that were observed below the barrier, and the per-
sistent challenging, also raises concerns about barrier breaches any time the barrier must be de-
energized for maintenance, and when barge vessels traverse the barrier.
- 3 -
Introduction
The Laurentian Great Lakes Basin has a long history of exotic species introductions
(Mills et al. 1993; Holeck et al. 2004). The Mississippi River Basin contains fewer invasive spe-
cies than the Great Lakes, however, it does have established populations of Bighead Carp (Hy-
pophthalmichthys nobilis) and Silver Carp (H. molitrix). These two species are of particular con-
cern because of their rapid population growth and planktivorous feeding ecology, which may
compete with native larval fishes and adult filter-feeding fish (Chick and Pegg 2001; Schrank et
al. 2003; Irons et al. 2007; Cooke and Hill 2010). Linking the Great Lakes and Mississippi River
Basins, and potentially serving as a conduit for inter-basin, invasive species exchange is the Chi-
cago Area Waterways System (CAWS).
The CAWS is a system of man-made canals and heavily modified rivers that now perma-
nently link the Great Lakes and Mississippi River Basins (Fig. 1). In 1910, upon completion of
the Chicago Sanitary and Ship Canal (CSSC), water flow in the Chicago River and the North
Shore Channel were permanently reversed, with water flowing from Lake Michigan to the Mis-
sissippi River Basin. This was done mainly for the purpose of diverting sewage away from Lake
Michigan, which is the source of drinking water for the city of Chicago, Illinois. Soon after the
completion of the CSSC, the Calumet-Saganashkee Channel (Cal-Sag Channel) was connected
to the CSSC, which reversed water flow from the Calumet, Grand Calumet, and Little Calumet
Rivers towards the Mississippi River Basin (Cain 1978; Moy et al. 2011).
- 4 -
Figure 1. Chicago Area Waterways System and location of electrical barriers.
Initially after the flow reversal, the amount of waste load from municipal and industrial
sources created a hypoxic zone that began in the CSSC, continued into the Des Plaines River,
and into the upper Illinois River. This “pollution barrier” rendered those systems devoid of fish
and prevented any inter-basin species exchange (Mills et al. 1966). Starting in the 1970’s, water
quality improved as a result of the Clean Water Act, enough so that fish began to inhabit the
CSSC (Moy et al. 2011). Today, approximately 70 different fish species occur in the CAWS
(Kevin Irons, Illinois DNR; personal communication). The water quality improvements have al-
so allowed invasive species such as zebra mussels (Driesenna polymorpha), White Perch (Moro-
ne americana), and Round Goby (Neogobius melanostomus), which had originally invaded the
- 5 -
Great Lakes, to expand their ranges into the CAWS and Illinois River (Stoeckel et al. 1996;
Steingraeber and Thiel 2000; Rasmussen 2002; Irons et al. 2002; 2006). Other potentially harm-
ful, non-native species such as Alewife (Alosa psuedoharengus), Oriental Weatherfish (Misgur-
nus anguillicaudatus), and Sea Lamprey (Petromyzon marinus) have been documented in the
CAWS (Retzer and Batten 2005).
In 1990, the U.S. Congress authorized the U.S. Army Corps of Engineers (USACE) to
study non-physical barriers in the CAWS in an effort to prevent the invasion of Round Goby into
the Mississippi Basin (Sparks et al. 2010). A group of scientists and managers formed the Dis-
persal Barrier Advisory Panel in 1996 and were tasked with considering various, non-physical,
fish dispersal barriers. Based on several factors such as cost, success likelihood, amount of envi-
ronmental impact, commercial availability, permit requirements, and effects on existing uses of
the CAWS, an electrical barrier was recommended as the best option for a non-physical fish bar-
rier (Moy et al. 2011).
The use of electricity to divert or prevent fish movement has origins in the early 20th cen-
tury with some of the first experiments taking place in the 1920s and 30s (McMillan 1928; Hart-
ley and Simpson 1967). Early electrical fish barriers consisted of metal electrodes spanning the
entire vertical water column of rivers, thus impeding fish movement via electrical repulsion, but
also creating physical barriers (McLain 1957; Hartley and Simpson 1967). In rivers where physi-
cal obstructions are undesirable, bottom-mounted electrode arrays have become more common
(Swink 1999; Moy et al. 2011). The effectiveness of electrical fish barriers have been evaluated
in controlled laboratory and field settings (Barwick and Miller 1996; Savino et al. 2001; Dawson
et al 2006; Holliman 2011) and at permanent barrier locations in small streams and canals
(Swink 1999; Verrill and Berry 1995; Maceina et al. 1999; Clarkson 2004). Effectiveness of
- 6 -
electrical barriers in controlled laboratory settings were evaluated via direct observation. Studies
in field settings relied on indirect assessment methods such as mark-recapture, and tagging, as
well as sampling above the barrier for the targeted species. Although the barriers in the above-
mentioned studies were largely effective, only Maceina (et al. 1999) found their electric barrier
to be 100% effective at inhibiting the movement of the targeted fish. Causes of barrier breach in
other studies included persistent challenging of the barrier by the fish (Barwick and Miller 1996;
Savino et al. 2001; Dawson et al. 2006; Holliman 2011), increased water flows (Verrill and Ber-
ry 1995), or unknown causes (Swink 1999). Clarkson (2004) extensively documented numerous
problems that arose at a barrier in an Arizona canal, mainly power outages, that resulted in grass
carp (Ctenopharyngodon idella) breach of the barrier.
In April 2002, an electric Demonstration Barrier was activated in the CSSC, however,
this was too late to stop the downstream dispersal of Round Goby, which were first found within
the Mississippi drainage in 1996 (Steingraeber and Thiel 2000; Sparks et al. 2010). This electri-
cal barrier is much different than any other past or present electrical barriers in several respects.
The section of the CSSC where the barriers are located, near Romeoville, Illinois, is 57-m wide
and 7.7-m deep making this electric barrier much larger than any other previous examples; flow
and conductivity vary greatly, and the canal is actively used for commercial and recreational ves-
sel navigation (Moy et al. 2010; Sparks et al. 2010).
Sparks et al. (2010) and Dettmers et al. (2005) were the first to directly test the effective-
ness of the Demonstration Barrier, which originally operated at 2 ms, 2Hz, < 0.39 V/cm (Holli-
man 2011). Sparks et al. (2010) released 130 Common Carp (Cyprinus carpio) with surgically-
implanted, combined radio-and-acoustic transmitters downstream of the barrier. Movements
were recorded from 2002 to 2006 and during that time one fish was able to breach the barrier on
- 7 -
April 3, 2003. This breach was later determined to have coincided with the passage of a barge
through the barrier, after a review of barrier electrical data, which shows distortions caused by
barges. This breach gave rise to speculation that the fish may have either been involuntarily en-
trained by the barge or that the barge may have distorted the electrical field enough that the fish
could have swam alongside the barge in an electrical void (Sparks et al. 2010). After the fish
moved upstream of the barrier, it ceased moving, indicating that the fish was either dead when it
breached, or died shortly after breaching. Despite uncertainty as to how the fish breached, the
barrier electrical parameters were increased to 5 ms, 3 Hz, < 0.39 V/cm, and later to 4 ms, 5 Hz,
0.39 V/cm in response to the fish passage upstream (Holliman 2011). The Demonstration Barrier
continues to operate today, but at a reduced capacity of about 0.31 V/cm because of its age
(McInerney et al. 2011).
During November 2003, Dettmers et al. (2005) passed encaged fish alongside a barge
through the Demonstration Barrier. Fish used were Catastomidae species, Morone species, and
Common Carp ranging in size from 170 – 580 mm TL. Dettmers et al. (2005) found that fish
took longer to show signs of affectedness from the barrier moving downstream than upstream
through the barrier. The effects of the electrical field were also delayed when fish swam along-
side conductive (steel) barge hulls compared to non-conductive (fiberglass) hulls. Some fish that
were towed along the steel-hulled barges were never incapacitated as they swam through the bar-
rier. Dettmers et al. (2005) attributed the delayed and non-incapacitations to a distortion of the
electrical field by the barges, which caused the field to warp towards the forward part of the
barge and created electrical voids immediately alongside and to the rear of the barge.
Following the Dettmers et al. (2005) study, design modifications were made to two larg-
er, additional electrical barriers, Barriers IIA and IIB, to account for the barge-induced electrical
- 8 -
warping. Barriers IIA and IIB were implemented in 2009 and 2011 respectively. The newer bar-
riers cover a much larger area than the Demonstration Barrier and are capable of generating elec-
trical fields of much higher intensity. These larger barriers also have parasitic structures, com-
posed of woven cable wire, placed upstream and downstream of them to prevent stray voltage
from reaching beyond the barrier “safety zone” (an area defined by the U.S. Coast Guard-Lake
Michigan Sector encompassing all of the barrier structures between river km 476.5 and 477.5 of
the CSSC). The two barriers consist of two downstream wide arrays that emit a weak electrical
field and two upstream narrow arrays that emit the maximum target voltage (Fig. 2; Holliman
2011). The purpose of this gradual increase in electrical intensity, moving from downstream to
upstream, is for fish to feel an increasingly unpleasant sensation, causing them to slow their
swimming speed and eventually stop. Otherwise, if fish abruptly encounter a narrow, high elec-
trical field, this may cause a panic response, in which the fish may continue to swim further into
the barrier until it breaches the barrier under its own momentum (Hartley and Simpson 1967).
Non-electrified spaces between the arrays were implemented with the intention of allowing fish
that are attempting to swim along a barge to move away from the barge in those areas (Moy et al.
2011).
After the completion of Barrier IIA in 2009, additional field testing was performed by
Sass and Ruebush (2010) and intensive laboratory work was performed by Holliman (2011) to
find the most effective barrier operating parameters. The main focus of this work was preventing
the upstream dispersal of Silver and Bighead Carp through the CSSC. Sass and Ruebush (2010)
tested the effectiveness of the Demonstration Barrier and Barrier IIA when both were operating
at 0.39 V/cm. Bluegill (Lepomis macrochirus), Common Carp, Golden Shiner (Notemigonus
crysoleucas), Green Sunfish (Lepomis cyanellus), Lepomis hybrids, Largemouth Bass (Microp-
- 9 -
terus salmoides), Orangespotted Sunfish (Lepomis humilis), Pumpkinseed (Lepomis gibbosus),
and White Sucker (Catostomus commersoni) ranging in size from 66 – 262 mm TL were at-
tached to tethers with floats, placed directly in the barriers, and responses to the barrier were di-
rectly observed and recorded. Only 3 out of 11 fish placed in the barriers operating at 0.39 V/cm
were incapacitated. However, later when the operating parameters of Barrier IIA were increased
to 6.5 ms, 15 Hz, 0.79 V/cm (0.79 V/cm), all 20 fish placed in the barrier were incapacitated.
Sass and Ruebush (2010) noted that smaller fish took longer to become incapacitated.
The operating parameters of Barrier IIA were increased to 0.79 V/cm in August 2009 as a
result of a pilot laboratory study performed by Holliman (2011). Using Silver Carp ranging in
size from 137-280 mm TL, Holliman (2011) found that at 0.79 V/cm, 100% of Silver Carp spec-
imens were incapacitated. Barrier IIB began operation in April 2011 at 0.79 V/cm. At the time of
our study, only one barrier (typically Barrier IIB) operated at a time along with the Demonstra-
tion Barrier (as of this writing both Barriers IIA and IIB concurrently operate along with the
Demonstration Barrier). Barrier IIB operated at 0.79 V/cm until November 2011, when parame-
ters were increased to 2.5 ms, 30 Hz, 0.91 V/cm (0.91 V/cm), which are the current operating
parameters as of this writing. The increase to 0.91 V/cm was in response to additional intensive
laboratory work done by Holliman (2011) on Bighead Carp that were 46-72 mm TL. Holliman
(2011) found that those parameters incapacitated 100% of small Bighead Carp that were exposed
to gradual increases in voltage in a Brett swim tunnel. However, those parameters were only
about 90% effective at preventing fish from swimming through an electrical barrier, set at the
current operating parameters of 0.91 V/cm, in a flowing runway that small Bighead Carp (48 –
82 mm TL) were allowed to challenge.
- 10 -
Given this history of studies that indicate potential entrainment of fish at the barrier was
an issue, Carterville FWCO undertook these studies to further understand potential issues with
entrainment at the barrier system under environmental conditions.
The behavior and abundances of wild fish that were present near the electric dispersal
barrier in the CSSC were recorded as well as the behavior of caged fish that were moved through
the barrier as part of this study. The behavior of specific fishes of interest that encounter electri-
cal barriers has been described in both laboratory (McMillan 1928; Hadderingh and Jansen 1990;
Savino et al. 2001; Dawson et al. 2006; Holliman 2011) and controlled field settings (Stewart
1981). Numerous studies exist on the use of electric barriers that are meant to divert fish away
from an undesirable area, such as a power plant intake (e.g. Taft 2000). However, to the best of
our knowledge, we are not aware of any other research that has focused on the behavior of estab-
lished, wild fish populations that encounter a permanent fish barrier designed to completely stop
upstream movements of fish. Godlewska et al. (2007) evaluated the abundances of wild fish in
front of an electric barrier using hydroacoustics, however, fish were recorded as “still” images,
and behavior could not be directly observed. We used a SONAR technology (described in detail
below) that was capable of recording target movements with video quality, and therefore record-
ed the fish’s interactions with the electric barrier. By using this technology we were able to easi-
ly quantify the interactions of both individual and schooled fish with the barrier. We also sam-
pled the electric barrier over a large area, from non-electrified water to the strongest part of the
barrier, throughout an entire year. The use of caged fish that are moved through an electrical bar-
rier is still a novel, experimental approach and the only other researchers to employ this method,
that we are aware of, were Dettmers et al. (2005) on the Demonstration Barrier within the CSSC.
- 11 -
Field work at the larger barriers (IIA and IIB) was performed in June 2011 and continued
after operating parameters were increased later that year. Because of the change in operating pa-
rameters in November 2011 to 0.91 V/cm, results from June to November 2011 are reported sep-
arately (when operating parameters were at 0.79 and 0.91 V/cm). Specific objectives at the pre-
vious operating parameters were to 1) describe the behavior of fish within various parts of the
electric barrier, 2) determine the abundances of fish within various parts of the electric barrier,
and 3) describe the behavior of encaged fish that are pulled through the barriers. Additional ob-
jectives after the current barrier operating parameters were initiated were to: 1) Determine the
number of fish interacting with the barrier at the peak voltage area, and 2) continue to describe
the behavior of encaged fish that are moved through the barriers. Initial caged-fish work consist-
ed solely of using aluminum-hull boats. However, later fiberglass-hull boats were also used in
order to contrast the effects that the metal-hull boat had on the barrier’s electrical field and
whether fish behavior was affected by any potential field distortion.
In addition to caged-fish trials at the electric barrier, on June 12, 2012, caged-fish trials
were performed within the Demonstration Barrier. These additional studies were performed after
a brief barrier outage in May 2012 raised concerns about how large fish interact with the Demon-
stration Barrier after breaching the primary barrier.
- 12 -
Materials and Methods
Fish behavior and abundance near the barriers
Fish abundances and behaviors within the electric barrier were evaluated using dual-
frequency identification sonar (DIDSON; Sound Metrics Corp., Bellevue, WA). The use of
DIDSON by fishery scientists to obtain size, behavior, and abundance data is becoming increas-
ingly common (Holmes et al. 2006; Boswell et al. 2008; Burwen et al. 2010; Becker et al. 2011).
A DIDSON has two available frequency modes: a low frequency of 1.0 MHz, which has a max-
imum viewing range of 40 m, but a lower target resolution, and a high frequency of 1.8 MHz that
has a lower viewing range of 12 m, but higher target resolution. The DIDSON software can pro-
cess up to seven frames per second, so rather than an acoustic “still” image, DIDSON provides,
near-video-quality, real-time imaging (Moursund et al. 2003). A DIDSON device is capable of
capturing consistent images regardless of turbidity and light levels (Moursund et al. 2003), giv-
ing this technology advantages to traditional underwater video camera viewing (Stoner 2004;
Mueller et al. 2006). A DIDSON is also capable of simultaneously imaging substrate, habitat
features, and fish within the same pulses. This makes images much more interpretable than other
SONAR methods (Boswell et al. 2008), especially when there is an interest in relating fish abun-
dance and behavior to habitat features.
DIDSON footage was recorded at eight separate sites (Fig. 2; Table 1) with ten cross-
channel sub-sites within each site (Fig. 3) for a total of 80 unique recordings during a sampling
period (see Table 1 for description of sites). The sites were chosen to represent the entire gradi-
ent of voltage along the electric barrier. In-water voltages begin to rise at the downstream para-
sitic structure (site 2) and gradually rise until it peaks between the two narrow, high-field arrays
- 13 -
(site 6). The voltage then abruptly decreases at the upstream parasitic structure (site 7; Fig. 2 [for
detailed maps of voltages throughout the electric barrier see McInerney et al. (2011)]).
Ten-minute recordings were made at each sub-site. The order in which to sample the sub-
sites was randomized with a random number generator. The DIDSON was operated in high-
frequency mode with a window length of 10 m starting 1.67 m in front of the transducer at a
frame rate of seven frames per second. The DIDSON unit was mounted to a heavy-duty pan and
tilt unit (Remote Ocean Systems, San Diego, CA) that was attached to a vertical metal pole. The
vertical pole was clamped onto a custom stanchion mount that was bolted to the gunwale of a
boat (Deep Development Corp., Sumas, WA). The DIDSON unit was deployed 1 m below the
water surface. The DIDSON unit was tilted at 0˚ for sub-sites A and J (in this case the 0˚ starting
orientation was considered to be when the DIDSON unit was parallel to the water surface), -25 ˚
for B and I, and -45 ˚ for C-H (Fig. 3). Environmental data (water velocity [m/s], dissolved oxy-
gen [mg/l], pH, conductivity [µs/cm], and temperature [˚C]) were recorded once during each re-
cording when the DIDSON sampled a site (roughly every ten minutes). These data were collect-
ed at a fixed area downstream of the barriers. Sampling took place continuously during daylight
hours with the only interruptions occurring when a vessel passed through the barrier. No record-
ings were made during vessel passages and we waited several minutes to resume sampling after a
vessel passage.
- 14 -
Figure 2. Schematic of the electric barrier within the CSSC where DIDSON images were record-ed. Note: drawing is not to scale and represents DIDSON recordings taken if Barrier IIB was operating. Sites were adjusted accordingly when Barrier IIA was in operation.
- 15 -
Figure 3. Cross-sectional illustration of the CSSC displaying ten sub-sites (A-J) across the canal where DIDSON footage was recorded within sites 1-8 from a north-oriented boat. Note: only one boat was used at a time to record one sub-site.
Table 1. Description of DIDSON recording sites.
After collection, DIDSON footage was reviewed in the laboratory. Two reviewers exam-
ined the footage simultaneously and were allowed to slow down, pause, and rewind footage in
order to best process all data. Individual fish (defined as a singular fish within the DIDSON
viewing cone that was not swimming in a similar manner as any nearby fish) were treated differ-
ently from schools/shoals of fish (defined as two or more fish swimming similarly nearby each
other and referred to as schools hereafter). Data were recorded separately for individual and
schooled fish based on the results of Holliman (2011) who found that small Bighead Carp did not
act independently when challenging an electrical barrier in a laboratory setting. Fish used by
Site Number DescriptionSite 1 Area downstream of all electrical structures where water-borne electricity is typically minimal.Site 2 Area immediately downstream of the first operating parasitic structure where water-borne electricity is typically minimalSite 3 Middle of first operating, downstream parasitic structureSite 4 Area immediately downstream of the first operating wide-array, low-field structureSite 5 Area immediately downstream of the second electrode bank of the wide-array structureSite 6 Area between the two narrow, high-field arrays where voltage is typically highestSite 7 Middle of first operating, upstream parasitic structureSite 8 Area upstream of all barrier structures where voltage is typically minimal
- 16 -
Holliman (2011) would often approach and challenge the barrier in pairs or as a group, or fish
would follow others into the electrical field.
Individual fish were measured with the DIDSON software measurement tool (cm) and
behaviors were recorded. Four different fish behaviors were recorded: “swimming,” “flight,”
“hovering,” and “probing.” “Swimming” was defined as steady, continuous movement by the
fish in one general direction. “Flight” was defined as a noticeable deviation from normal swim-
ming behavior including bursts of speed or rapid, erratic changes in direction. “Hovering” was
defined as fish maintaining position in one spot for at least five seconds. “Probing” was defined
as direct and persistent back and forth and/or up and down movement where forward progress
seemed to be inhibited along an invisible plane. All observed behaviors were recorded for an in-
dividual fish or schools of fish. For example, if a fish showed hovering behavior and then prob-
ing behavior, both were recorded. When schools of fish came into the DIDSON view, the lengths
of five randomly selected fish were recorded and averaged. The same behavioral data that were
recorded for individual fish were also recorded for schools. Each fish that entered the DIDSON
viewing area was treated as a separate fish. Therefore, the possibility exists that if an individual
fish or school of fish was swimming back and forth through the DIDSON viewing area that those
fish would be counted multiple times. Thus, fish abundance estimates may be elevated at all
sites.
Only one season’s worth of DIDSON footage was recorded, using the standard method of
sampling 80 different subsites, when the barrier was operating at 0.79 V/cm. Two separate meth-
ods of recording fish using the DIDSON unit were employed throughout 2012 after the barrier
operating parameters were increased to 0.91 V/cm. One which evaluated fish behavior and abun-
dance throughout the entire electric barrier, throughout the year (standard sampling; described
- 17 -
above) and another which focused on fish abundance and behavior within the zone of ultimate
field strength (concentrated sampling; Site 6 in Fig. 2). The concentrated DIDSON footage fo-
cused on the western wall of the canal within the zone of ultimate field strength. In the standard
sampling method, this area would be referred to as Site 6, Section A (Figs. 2 and 3). This site
was chosen for concentrated evaluation because earlier work at the previous operating parame-
ters had revealed that wild fish had been present at that location and interacted with the barrier
mostly at the water surface, near the canal wall. Environmental data (dissolved oxygen [mg/l],
pH, conductivity [µs/cm], and temperature [˚C]) were recorded once during each DIDSON re-
cording interval (during both standard and concentrated methods) at a fixed site downstream of
the barriers. Water velocity data, for the times that DIDSON sampling took place, were obtained
from a U.S. Geologic Survey water gage located in the CSSC 10.5 km upstream of the barrier
safety zone (http://waterdata.usgs.gov/il/nwis/uv/?site_no=05536890).
Caged-fish trials Gizzard Shad (Dorosoma cepedianum) was used as a surrogate species to emulate small
Asian carp attempting to traverse the barrier. Asian carp were not used in the caged-fish trials
because of escapement concerns. The gizzard shad was chosen as a surrogate species because its
body morphology and habitat preferences are similar to Asian carp and was locally abundant in
the CSSC so as to minimize potential collection handling stress. Clupeidae species such as giz-
zard shad are generally known to be sensitive to handling stress (e.g. Smith et al. 2009), likely
more so than Asian carp. Because of the sensitivity of Gizzard Shad compared to other potential
surrogates to Asian carp, we feel that the results obtained are likely conservative.
Gizzard Shad that were used in the caged-fish trials were collected via cast netting during
fall 2011 and winter 2011. Later, when fish became more difficult to locate, we collected Giz-
- 18 -
zard Shad via electrofishing for the remainder of the trials. All fish were collected on the same
day that the trials took place. Although collection of the Gizzard Shad used in the trials was ran-
dom in nature, large fish (> 300 mm TL) were not targeted. After the fish were collected, the
smallest fish available were used given that small fish are thought to pose the greatest threat to
breaching the barrier (Holliman 2011; Parker et al. 2013). All fish were measured after the trials
concluded as part of efforts to reduce handling stress.
The cage used in the trials had a non-conductive PVC frame (160-cm L x 58-cm W x 89-
cm D) with 0.95-cm bar monofilament mesh. The cage was secured alongside a boat using cus-
tom mounts (figure 4). During the first evaluation in the summer of 2011, a DIDSON unit,
mounted on the opposite side of the boat, was used to record fish behavior in the cage. However,
after the first evaluation with the DIDSON unit, we found that water clarity and visibility al-
lowed us to record fish behavior with a camcorder mounted above the cage. Using a camcorder
allowed for faster set up time in the field and allowed for easier navigation of the boat without
the DIDSON in the water. Later comparisons of camcorder and DIDSON recordings revealed no
differences in reviewer’s ability to interpret fish behavior.
- 19 -
Figure 4. Cage mounted on the side of an aluminum boat.
Figure 5. Illustration depicting the cage and DIDSON mounting schemes on the north-oriented research boat in the CSSC and the two parts of the canal that the boat travelled. Note: only a single boat was used for caged-fish trials.
- 20 -
Five gizzard shad were placed in the cage and moved through one of the following three
115-m sections of the CSSC moving from south to north (upstream): (a) through the mid-channel
of the CSSC over the entire array of electrical barrier structures, (b) along the western canal wall
of the CSSC, over the entire array of electrical barrier structures, or (c) a control area through the
mid-channel of the CSSC, in non-electrified water (Fig. 6). The upper end of the control trial run
was approximately 100-m downstream of all electrical barrier structures and outside of any elec-
trical influences. For the runs that were performed through the electric barrier, the starting point
was in an area of non-electrified water. Ten trial runs were performed within each section of the
canal per experimental event.
Fish were given 1 min before the trial began to acclimate to the cage in non-electrified
water below the electric barrier. After acclimation, we began recording the fish. When the DID-
SON was used, a stop watch was started at the same time as the DIDSON. Times were then rec-
orded when the caged fish passed by nine different observation points (Fig. 6; Table 2). These
times were later used by workers to record fish behavior at the designated sites. When the cam-
corder was used, the observation point numbers were announced into the camcorder microphone.
These verbal cues were later used in the lab to record fish behavior at the designated points.
Trial evaluations using gizzard shad occurred once during the summer of 2011 when the
barrier operating parameters were 0.79 V/cm. All other evaluations occurred at the current oper-
ating parameters of 0.91 V/cm. Caged-fish evaluations occurred once during the summer and fall
of 2011, and winter 2012 using solely an aluminum-hull boat. In the spring and summer of 2012,
caged fish were moved through the electrical barriers using both an aluminum-hull and a fiber-
glass-hull boat. The trials using both aluminum and fiberglass-hull boats were performed within
a two-week period during the spring and summer to reduce temporal variability. Boat speeds
- 21 -
were adjusted continuously throughout the trial runs depending on the swimming speed of the
fish used (typically 1.6-3.2 km/h). Fish that became incapacitated after moving through the barri-
er were observed beyond the barriers to assess recovery rate.
- 22 -
Figure 6. Schematic of the electric barrier within the CSSC where caged-fish were moved through from downstream to upstream. Observation points used in the control runs were the same dis-tances apart but outside of the barrier area. Observation points are where behaviors were later recorded. Note: drawing is not to scale and represents caged-fish observation points if only Barrier IIB was operat-ing. Sites were adjusted accordingly if Barrier IIA was operating.
- 23 -
Table 2. Description of caged-fish observation points. Recordings of caged fish were reviewed in the laboratory, separately, by two different
workers. At each observation point, one of four different behaviors was recorded for each fish:
“swimming,” defined as normal movement of the fish keeping pace with the moving boat,
“flight,” defined as a deviation from normal swimming behavior including sudden bursts of rapid
swimming, multiple, rapid changes in position, and non-directional swimming, “floundering,”
defined as a fish struggling to not become impinged on the downstream end of the cage, floun-
dering behavior includes thrashing around against the downstream end of the cage or pushing
itself off of the cage momentarily before being swept back down against the cage, and “incapaci-
tation,” which was used when all movement ceased and the fish became impinged against the
downstream end of the cage. If the two reviewers assigned conflicting behaviors for a fish at an
observation point a third reviewer assigned a behavior. If the third assigned behavior matched
one of the previous two assigned behaviors, that behavior was used. If three different behaviors
were assigned, for a fish at a particular observation point, no behavioral datum was recorded.
Observation point Description1 Area immediately below the first operating parasitic structure where voltage is minimal2 Middle of first operating parasitic structure3 Mid-point between parasitic structure and first wide array structure4 Middle of downstream wide-array structure5 Mid-point between two wide-array structures6 Middle of upstream wide-array structure7 Middle of downstream narrow-array structure8 Mid-point between two narrow-array structures where voltage is highest9 Middle of upstream narrow array structure
- 24 -
Demonstration barrier caged-fish trials
On June 12, 2012, nine caged-fish trials were performed within the Demonstration Barri-
er using 11 Common Carp (391-673 mm TL) and 2 Freshwater Drum (Aplodinotus grunniens;
371-384 mm TL). Four observation points were designated for verbal cues to be announced
while the camcorder recorded behavior. The first observation point was located 39 m south of the
Demonstration Barrier, in non-electrified water. Observation points 2-4 were located at the
downstream, middle, and upstream locations of the Demonstration Barrier, which covers 18 m of
canal length. All trials were performed using a fiberglass boat.
- 25 -
Statistical Analyses
DIDSON sampling data analyses
Prior to analysis of count data from the DIDSON recordings, the data needed to be stand-
ardized to account for unequal volumes of water being ensonified by the DIDSON cone. The
DIDSON ensonified the same volume of water at all of the site sub-sections with the exceptions
of “A” and “J.” The viewing cone at those sections intersected with the water surface. In geomet-
ric terms, the volume of the truncated DIDSON ensonification cone that was lost is referred to as
a vertical slice. The vertical slice loss was calculated using mathematical integration by section
of the circular cross section of the truncated cone from the cone bottom to the intersection of the
water surface. Assuming a distance of eight meters from the DIDSON unit to the canal wall, a
loss of 11% of the DIDSON viewing cone was determined. Therefore, the numbers of fish in
sections A and J received a weighting of 1.11 in all regression models in order to account for the
vertical slice loss; all other sections received a weighting of 1.
The probabilities of individual fish and schooled fish presence were compared across lo-
cations (i.e., sites or sections). Specifically, we used binary logistic regression models to deter-
mine whether the probability of individual fish presence was different based on the location
within the electric barrier (i.e., site or section); the logit-link function was used to transform loca-
tion-specific probabilities to the natural log of the odds ratio of presence versus absence. Post-
hoc odds ratios were used to examine which locations were responsible for significant effects.
Odds ratios that were significantly different from one indicate that the probability of individual
or school presence is greater in the locations or seasons being compared (Agresti 2007). To more
closely examine the spatial differences in section presence of individual or schooled fish within
each site, we included both the main effects of site, section, and their interaction; this resulted in
- 26 -
several site and section combinations that had probabilistic estimates of zero or one, thereby in-
hibiting the number of comparisons on the basis of quasi-complete separation of data points. In
other words, too few observations were available to make statistical inferences outside of visual
interpretations of estimated probabilities for this model. In all models, beam volume was used as
a weighting factor as described above. The effect that fish size has on determining which site a
fish was present in was evaluated using a logit model. Wald χ2 statistics were used to determine
whether size had an overall effect on fish distributions throughout the electric barrier. Similar to
Becker et al. (2011), one behavioral category was often dominant at each site, leaving insuffi-
cient data to perform simple statistical analyses, therefore behavioral data are described.
Durbin-Watson tests were performed on the environmental variables that were recorded
during each DIDSON run, which found that those data were highly autocorrelated. Therefore
means of those data were taken for the entire sampling period and separate simple linear regres-
sions were performed with mean loge(n+1)-transformed individual fish abundances and number
of fish schools against velocity, dissolved oxygen, temperature, and conductivity. For the regres-
sion analyses, alpha was Bonferroni-corrected for multiple comparisons (5 for both individual
and number of fish schools), therefore, alpha was set at 0.01.
Stepwise logistic regression was used to evaluate factors influencing fish presence within
the zone of ultimate field strength. Independent variables tested in the model were water velocity
(m/s), dissolved oxygen (mg/l), pH, conductivity (µs/cm), and temperature (˚C). The environ-
mental data were recorded each time a DIDSON recording took place (every ten minutes). Dur-
bin-Watson tests revealed that the environmental data were autocorrelated within each day that
the recordings were taken. Therefore, the mean values of the environmental data were taken for
- 27 -
each day. Separate stepwise logistic regressions were performed modeling the presence (scored a
one)/absence (scored a zero) of schooled and individual fish during each day.
Caged-fish data analysis Binary logistic regression was used to model the probability of fish incapacitation as a
function of distance into the electric barrier from our predetermined starting point (Agresti
2007). Fish that immediately became impinged on the back of the cage when the boat first start-
ed moving through non-electrified water, presumably because of previous capture and handling
stress, were omitted from analyses, which was 19% of the fish used. By modeling the binary re-
sponse of each fish randomly assigned to a treatment (control, mid-channel, west wall), each fish
was considered to represent a single Bernoulli trial, or replicate, in which only two outcomes
were possible (Breen and Ruetz 2006; Agresti 2007; Gotelli and Ellison 2004). However, if the
assumption that each fish represents a Bernoulli trial is not approximately true slope and inter-
cept estimates would not be affected, yet the probability of a type I error would be inflated
(Breen and Ruetz 2006; Agresti 2007). Initially, the logistic regressions that were performed on
the spring, fiberglass-hull boat results had quasi-complete separation of data points (i.e., a specif-
ic distance at which all fish went from not being incapacitated to being incapacitated). For con-
vergence and comparative purposes a score of one was artificially added to each observation dis-
tance (Allison 2008).
The distance at which 50% of gizzard shad were incapacitated (median incapacitation
distance, hereafter) was estimated from the slope and intercept estimates from each logistic re-
gression function. To evaluate the effect of temperature, conductivity, and fish size on median
incapacitation distance, a principal components analysis (PCA) was used to reduce the dimen-
sionality of the correlated independent variables into a single value, or PCA score, for each trial
- 28 -
period. This analysis was only performed for the present barrier voltage in which an entire year
of data was available. Median incapacitation distances were log-transformed to homogenize var-
iances and a linear regression was used to examine the relationship between PC 1 scores and me-
dian incapacitation distance. A Pearson correlation between the PC scores 1 and 2 and the inde-
pendent variables was performed in order to assess relationships between the individual variables
and the PC scores. A lack of within-season replication precluded analyses relative to the effects
of canal location (canal wall or mid-channel) and boat hull material, so results are described.
- 29 -
Results
Former Operating Parameters (0.79 V/cm) Fish behavior and abundance near the barriers
During the summer of 2011, 750 individual fish and 570 different schools of fish were
observed within and around the electric barrier. The probability of individual fish presence dif-
fered among sites (Wald χ2 = 71.11; df = 7; P < 0.001), as did the probability of schooled fish
presence (Wald χ2 = 51.57; df = 7; P < 0.001), with the highest probabilities being observed at
sites 1-4 for both individual fish (Fig. 7) and schooled fish (Fig. 8), which are below the low-
field electrical arrays. However, some fish were observed further into the electric barrier, includ-
ing site six, which is between the narrow arrays (Fig. 7 and Fig. 8) that contained the ultimate
field strength of 0.79 V/cm.
The probability of individual fish presence was not different among sections when pooled
across sites (Wald χ2 = 12.20; df = 9; P = 0.202), with a 0.92 probability of presence across all
sections (95% CI = 0.90 – 0.93). Schooled fish probability of presence on the other hand did vary
among sections pooled across sites (Wald χ2 = 45.11; df = 9; P < 0.001), with higher probabili-
ties of presence closer to the canal walls (Fig. 9). Within each site, both individual fish and
schools of fish (Fig. 10) utilized the majority of the sections in sites one through four. However,
within the sites that were in the highest-voltage water and upstream of the barriers, the fish had a
higher probability of being present near the canal walls, particularly the east wall.
- 30 -
Figure 7. Probability of individual fish presence at sites 1-8 within the electric dispersal barrier in summer 2011 (pooled across sections). Bars with the same letters indicate statistically similar odds ratio estimates between sites (P > 0.05).
Site
1 2 3 4 5 6 7 8
Pro
babi
lity
of In
divi
dual
Fis
h P
rese
nce
(95%
CI)
0.0
0.2
0.4
0.6
0.8
1.0a a
bb
cdc
dd
- 31 -
Figure 8. Probability of schooled fish presence at sites 1-8 within the electric dispersal barrier in summer 2011 (pooled across sections). Bars with the same letters indicate statistically similar odds ratio estimates between sites (P > 0.05).
Site
1 2 3 4 5 6 7 8
Pro
babi
lity
of S
choo
led
Fish
Pre
senc
e (9
5% C
I)
0.0
0.2
0.4
0.6
0.8
1.0a
a a a
b bb
b
- 32 -
Figure 9. Probability of schooled fish presence at sections A-J within the electric dispersal barrier in summer 2011 (pooled across sites). Bars with the same letters indicate statistically similar odds ratio estimates between sections (P > 0.05).
Section
A B C D E F G H I J
Pro
babi
lity
of in
divi
dual
fish
pre
senc
e (9
5% C
I)
0.0
0.2
0.4
0.6
0.8
1.0 aa ab aabbc bc
c cc
- 33 -
Section
A B C D E F G H I J
Site
1
2
3
4
5
6
7
8 Probability
A B C D E F G H I J
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Individual fish Schooled fish
Figure 10. Probabilistic estimates of individual fish and schooled fish presence plotted across sites and sections within the electric dispersal barrier from summer 2011.
We found that for both individual (Fig. 11; Wald χ2 = 134.63, P < 0.0001) and schooled
fish (Fig. 12; Wald χ2 = 83.37, P < 0.0001), size was a significant factor in determining which
site fish were present in. Larger fish had the greatest probability of being present at sites one
through four. At sites five through eight, small fish had the highest probabilities of being present.
- 34 -
Figure 11. Probability models for individual fish presence based on size at each site. Black lines represent predicted probabilities and gray lines represent 95% confidence intervals.
- 35 -
Figure 12. Probability models for schooled fish presence based on size at each site. Black lines represent predicted probabilities and gray lines represent 95% confidence intervals.
- 36 -
For both individual fish and schools of fish, swimming was the dominant behavior that
was observed at all sites with the electric barrier. Hovering behavior increased when fish were
closer to the barriers, and peaked at sites 5 and 6, most likely because those fish were between
barrier arrays. Site 4 was the only site where fish displayed all four possible behaviors (fig. 13;
14). Fish that encountered the barrier were often observed swimming both up and down and
sideways along the electrical field, fish were also observed swimming into the electrical field and
then returning downstream. None of the linear regressions relating numbers of individual fish
and schools of fish to environmental variables were significant at alpha = 0.01 given the low
power of detection (p ≥ 0.017)
- 37 -
Sites
1 2 3 4 5 6 7 8
%
0.0
0.2
0.4
0.6
0.8
1.0
SwimmingFlightHoveringProbing
Figure 13. Percentages of different behaviors observed by individual fish at the different sites within the electric barrier in the CSSC.
- 38 -
Sites
1 2 3 4 5 6 7 8
%
0.0
0.2
0.4
0.6
0.8
1.0
1.2
SwimmingFlightHoveringProbing
Figure 14. Percentages of different behaviors observed by schools of fish at the different sites within the electric barrier in the CSSC.
- 39 -
Former Operating Parameters (0.79 V/cm) Caged-fish trials Of the 270 fish that were pulled through the electric barrier, 3% did not become
incapacitated (n=4 through the mid-channel and n=4 along the west canal wall). One of the fish
that was moved through the mid-channel exhibited erratic swimming, in which it was briefly im-
pinged on the back of the cage (but still moving); the other seven fish swam the entire time while
moving through the barriers. The estimated slopes and intercepts for the logistic regression mod-
els for all run locations were significant (p < 0.001) at the lower operating parameters. The lo-
gistic regression model for the control runs showed a low probability of fish becoming incapaci-
tated as a result of being forced to swim along a boat while caged and the proportion of variabil-
ity explained was very low (generalized R2 = 0.38, slope = -3.69 [SE ± 0.357], intercept = 0.01
[SE ± 0.004]; Fig. 15). The logistic regression models for the mid-channel (generalized R2 =
0.74, slope = -4.09 [SE ± 0.265], intercept = 0.06 [SE ± 0.003]) and west wall (generalized R2 =
0.75, slope = -5.008 [SE ±0.328], intercept = 0.07[SE ± 0.004]) runs had higher proportions of
variability explained relative to the control location given that distance was a better predictor for
incapacitation. The median incapacitation distances were 66 m in the mid-channel and 73 m
along the west wall (Fig. 15). The average sizes of fish (mm ± SE) used in the control, mid-
channel, and canal wall trials were 126.7 (5.72), 123.7 (5.73), and 111.3 (3.03) respectively. The
average water conductivity was 0.74 mS/cm (SE ± 0.003), and average water temperature was
24.2 °C (SE ± 0.105).
Swimming and incapacitation were the most common behaviors recorded. Swimming
was the most common recorded behavior for the control runs (fig. 16). Swimming behavior de-
creased steadily as fish were moved further into the electric barrier in both the mid-channel and
- 40 -
along the west wall. Flight and floundering behavior, while low, occurred mostly at sites one
through five before fish became incapacitated (figs. 17;18).
- 41 -
Figure 15. Predicted incapacitation probabilities of encaged gizzard shad as a function of distance at different locations with the barrier operating at 0.79 V/cm. Lines extending from x and y axes denote predicted point at which half of the fish be-came incapacitated (mid-channel and west wall only). Points represent the means (± SE) of observations at site distances, but the model is based on actual observations (0 or 1). Shaded areas represent locations of wide and narrow barrier arrays.
- 42 -
Figure 16. Percentages of different recorded behaviors for caged fish at each observation point in the control area.
- 43 -
Figure 17. Percentages of different recorded behaviors for caged fish at each observation point in the mid-channel of the electric barrier.
- 44 -
Figure 18. Percentages of different recorded behaviors for caged fish at each observation point along the west wall of the electric barrier.
Current Operating Parameters (0.91 V/cm) Fish behavior and abundance near the barriers
Throughout the winter, spring, summer, and fall of 2012 we respectively sampled 74, 41,
45, and 123 individual fish and 46, 46, 215, and 246 schools of fish. The probability of individu-
al fish presence differed among sites (Wald χ2 = 18.75; df = 7; P = 0.009) and seasons (Wald χ2
= 25.96; df = 3; P < 0.001), yet site-specific probabilities of individual fish presence were not
consistent across seasons (Wald χ2 = 39.77, df = 21, P = 0.008). In general, the odds of individu-
al fish presence across all sites was greater in the fall compared to the summer (Wald χ2 = 25.96;
- 45 -
df = 1; P < 0.001), but were similar among all other seasons (Wald χ2 ≤ 0.001; df = 1; P ≥
0.974). Within-season differences of the probability of individual fish presence among sites oc-
curred in the winter (Wald χ2 = 22.47, df = 7, P = 0.002), summer (Wald χ2 = 22.12, df = 7, P =
0.002), and spring (Wald χ2 = 18.21, df = 7, P = 0.011), but were not different in the fall (Wald
χ2 = 11.77, df = 7, P = 0.108; Figure 19). Probabilities of observing schooled fish at specific
sites also differed seasonally (Wald χ2 = 60.59, df = 21, P < 0.001); the main effects of site
(Wald χ2 = 0.0004; df = 7; P = 1.00) and season (Wald χ2 = 0.0004; df = 7; P = 1.00) were not
significant. Probabilistic estimates of fish school presence were different among sites in the
summer (Wald χ2 = 37.94, df = 7, P < 0.001) and fall (Wald χ2 = 31.42, df = 7, P < 0.001), but
were not different in the winter (Wald χ2 = 4.47, df = 7, P = 0.724) or spring (Wald χ2 = 8.90, df
= 7, P = 0.260; Fig. 20). Both the individual and schooled fish were largely absent from sites 4-6
during the winter and spring periods. However, fish began to penetrate further into the barriers in
the summer and by the fall, with both individual and schooled fish observed as far into the barri-
er as site six (Figs. 23 and 24) which contains the ultimate field strength of 0.91V/cm.
- 46 -
Winter 2012
1 2 3 4 5 6 7 80.0
0.2
0.4
0.6
0.8
1.0 Spring 2012
1 2 3 4 5 6 7 80.0
0.2
0.4
0.6
0.8
1.0
Summer 2012
1 2 3 4 5 6 7 8
Pro
babi
lity
of In
divid
ual F
ish
Pre
senc
e (9
5% C
I)
0.0
0.2
0.4
0.6
0.8
1.0Fall 2012
Site
1 2 3 4 5 6 7 80.0
0.2
0.4
0.6
0.8
1.0
aab
ab
bcbc
c
* * * *
a
aa
b
b b
aab
a
ab
bb b b
aa a
aa
aa
a
Figure 19. Season-specific probabilities of individual fish presence at sites 1-8 within the electric dispersal barrier throughout 2012 (pooled across sections). Bars with similar letters indicate statistically similar odds ratio estimates between sites within a season (P > 0.05). An asterisks (*) indicates that indi-vidual fish were never observed at these sites such that no valid comparisons were possible.
- 47 -
Winter 2012
1 2 3 4 5 6 7 80.0
0.2
0.4
0.6
0.8
1.0 Spring 2012
1 2 3 4 5 6 7 80.0
0.2
0.4
0.6
0.8
1.0
Summer 2012
1 2 3 4 5 6 7 8
Pro
babi
lity
of S
choo
led
Fish
Pre
senc
e (9
5% C
I)
0.0
0.2
0.4
0.6
0.8
1.0Fall 2012
Site
1 2 3 4 5 6 7 80.0
0.2
0.4
0.6
0.8
1.0
aa
a
* * * *
a
aa
aa
*
aa a
c
a
*
c
b
b b a
c
b
c
ab b
* **
Figure 20. Season-specific probabilities of schooled fish presence at sites 1-8 within the electric dispersal barrier throughout 2012 (pooled across sections). Bars with similar letters indicate statistically similar odds ratio estimates between sites within a season (P > 0.05). An asterisks (*) indicates that schooled fish were never observed at these sites such that no valid comparisons were possible.
We found that when comparing fish presence among all sites, that schooled fish exhibited
differential use of sections across seasons (Wald χ2 = 45.82, df = 27, P = 0.013). Within-season
probabilities of schooled fish presence differed among sections in the winter (Wald χ2 = 19.70; df
= 9; P = 0.020) and fall (Wald χ2 = 28.47; df = 9; P < 0.001), but was not different in the spring
(Wald χ2 = 12.18; df = 9; P = 0.204) or summer (Wald χ2 = 11.00; df = 9; P = 0.276; Figure 21).
Individual fish also exhibited differential use of sections across seasons (Wald χ2 = 55.65, df =
27, P = 0.001). Within-season probabilities of individual fish presence differed among sections
- 48 -
in the winter (Wald χ2 = 41.10; df = 9; P < 0.001) and fall (Wald χ2 = 20.65; df = 9; P = 0.014),
but was not different in the spring (Wald χ2 = 10.46; df = 9; P = 0.314) or summer (Wald χ2 =
10.03; df = 9; P = 0.3484; Fig. 22). Within sites 1-3 and site 8, both individual fish (Fig. 23) and
schools of fish (Fig. 24) utilized the majority of the sections. However, within the other sites that
were closer to the electrified portions of the barrier, the fish had a higher probability of being
present near the canal walls.
Winter 2012
A B C D E F G H I J0.0
0.2
0.4
0.6
0.8
1.0 Spring 2012
A B C D E F G H I J0.0
0.2
0.4
0.6
0.8
1.0
Summer 2012
A B C D E F G H I J
Pro
babi
lity
of S
choo
led
Fish
Pre
senc
e (9
5% C
I)
0.0
0.2
0.4
0.6
0.8
1.0Fall 2012
Section
A B C D E F G H I J0.0
0.2
0.4
0.6
0.8
1.0
** * *
a
aaa
a
a
a
a
aa
aaa
aaaaa
aababc
bc
ab
c cc
a ab ab b
c
ba a a
Figure 21. Season-specific probabilities of schooled fish presence at sections A-J within the elec-tric dispersal barrier throughout 2012 (pooled across sites). Bars with similar letters indicate statistically similar odds ratio estimates between sections within a season (P > 0.05). An asterisks (*) indicates that schooled fish were never observed at these sections such that no valid comparisons were possible.
- 49 -
Winter 2012
A B C D E F G H I J0.0
0.2
0.4
0.6
0.8
1.0 Spring 2012
A B C D E F G H I J0.0
0.2
0.4
0.6
0.8
1.0
Summer 2012
A B C D E F G H I J
Pro
babi
lity
of In
divid
ual F
ish
Pre
senc
e (9
5% C
I)
0.0
0.2
0.4
0.6
0.8
1.0Fall 2012
Section
A B C D E F G H I J0.0
0.2
0.4
0.6
0.8
1.0
aba
b
b bbb
bb
aa
aaa
aa
a
a
aaa
aaaa
aa
a
a
bc bc
ab a
bc bcc
ab
**
*
Figure 22. Season-specific probabilities of individual fish presence at sections A-J within the electric dispersal barrier throughout 2012 (pooled across sites). Bars with similar letters indicate statisti-cally similar odds ratio estimates between sections within a season (P > 0.05). An asterisks (*) indicates that individual fish were never observed at these sections such that no valid comparisons were possible.
- 50 -
Winter 2012
A B C D E F G H I J1
2
3
4
5
6
7
8
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Spring 2012
A B C D E F G H I J
Summer 2012
A B C D E F G H I J
Site
1
2
3
4
5
6
7
8Fall 2012
Section
A B C D E F G H I J
Probability
Figure 23. Probabilistic estimates of individual fish presence plotted across sites and sections within the electric dispersal barrier for each season within 2012. The lack of replicated observation events in spring, summer, and fall 2012 resulted in site and section probabilistic estimates to be either 0 or 1.
- 51 -
Winter 2012
A B C D E F G H I J1
2
3
4
5
6
7
8
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Spring 2012
A B C D E F G H I J
Summer 2012
A B C D E F G H I J
Site
1
2
3
4
5
6
7
8Fall 2012
Section
A B C D E F G H I J
Probability
Figure 24. Probabilistic estimates of schooled fish presence plotted across sites and sections within the electric dispersal barrier for each season within 2012. The lack of replicated observation events in spring, summer, and fall 2012 resulted in site and section probabilistic estimates to be either 0 or 1.
- 52 -
We found that size was only a significant factor in determining which sites fish were pre-
sent in for individual fish in the fall season (Fig. 25; Wald χ2 = 18.45, df = 7, P = 0.010) and
schooled fish in the summer season (Fig. 26; Wald χ2 = 33.20, df = 6, P < 0.0001) and did not
have an effect in other seasons for individuals or schools. Larger fish had the greatest probability
of being present at sites 1-3. At sites 4-8, small fish had higher probabilities of being present. For
both individual fish and schools of fish, in all seasons, swimming was the dominant behavior that
was observed at all sites within the electric barrier. Hovering behavior increased when fish were
closer to the barriers (Fig. 27; 28).
- 53 -
Figure 25. Probability models for fall season, individual fish presence based on size at each site. Black lines represent predicted probabilities and gray lines represent 95% confidence intervals.
- 54 -
Figure 26. Probability models for summer season, schooled fish presence based on size at each site. Black lines represent predicted probabilities and gray lines represent 95% confidence intervals. Note: no information is presented for site six because no fish schools were present at that time.
- 55 -
Figure 27. Proportions of different behaviors observed by individual fish at the different sites within the electric barrier in the CSSC.
- 56 -
Figure 28. Proportions of different behaviors observed by schooled fish at the different sites within the electric barrier in the CSSC.
- 57 -
The concentrated DIDSON sampling within the zone of ultimate field strength revealed a
low amount of fish activity in that area during the spring with only 16 individuals and 1 school of
fish observed. However, in the summer, fish activity in that area increased with 143 individuals
and 238 schools of fish observed. Stepwise logistic regression models revealed that dissolved
oxygen concentrations and temperature were the most informative variables explaining pres-
ence/absence of schooled fish within the zone of ultimate field strength (Wald χ2 = 12.37, df = 2,
P = 0.0021; Table 3). Dissolved oxygen concentration alone was most informative when explain-
ing individual fish presence/absence (Wald χ2 = 6.46, df = 1, P = 0.0110; Table 4). The average
size of both schooled (Fig. 29A) and individual (30A) fish that were present within the zone of
ultimate field strength decreased from the early spring to late summer.
Table 3. Summary of stepwise logistic regression models used to estimate the probability of individual and schooled fish presence/absence at the zone of ultimate field strength.
Variable Parameter Estimate Wald χ2 df P
Intercept -43.88 10.67 1 0.001Dissolved oxygen 5.74 9.88 1 0.002
Temperature 0.96 11 1 0.001
Intercept -24.24 5.64 1 0.018Dissolved oxygen 7.96 6.46 1 0.011
Schooled fish model
Individual fish model
- 58 -
Figure 29. Mean sizes of schooled fish present at the zone of ultimate field strength (A) and mean dissolved oxygen concentrations (B) and temperatures (C). Each bar represents a single day of sampling in April, May, July, and September 2012. Hatch marks in dissolved oxygen and temperature bars indicate days in which a school of fish was present in the zone of ultimate field strength.
Figure 30. Mean sizes of individual fish present at the zone of ultimate field strength (A) and mean dissolved oxygen concentrations. Each bar represents a single day of sampling in April, May, July, and September 2012. Hatch marks in dissolved oxygen bars indicate days in which indi-vidual of fish were present in the zone of ultimate field strength.
- 59 -
Flight, hovering, and probing behaviors were observed much more during the concentrated DID-
SON sampling within the zone of ultimate field strength, although swimming was still the most
dominant behavior (Fig. 31). Fish that encountered the barrier were often observed swimming
both up and down and sideways along the electrical field, fish were also observed swimming into
the electrical field and then returning downstream. The fish that were observed penetrating the
furthest upstream into the zone of ultimate field strength attempted to move further upstream by
swimming mostly perpendicular to the canal flow, but also moving slightly upstream. As they
swam, the fish changed the direction of perpendicular movement rapidly. This behavior was in-
terpreted to be body-voltage minimizing behavior.
Figure 31. Proportions of different behaviors observed by individual and schooled fish during the spring and summer of 2012 at the zone of ultimate field strength in the barrier.
- 60 -
Current Operating Parameters (0.91 V/cm) Caged-fish trials All fish that were moved through the barriers, at the current operating parameters of 0.91
V/cm, were incapacitated at some point. The probability of incapacitation was strongly related
with distance into the electric barrier as indicated by significant slope values for all but two tri-
als, one of which was a control (Table 4). The distance into the electric barrier in which fish be-
came incapacitated varied widely, however, depending on season, which part of the canal fish
were moved through (mid-channel or along west wall), or the type of boat used (fiberglass or
aluminum hull) (Fig. 32). Fish that were moved through the barriers in the winter and spring be-
came incapacitated the soonest, whereas during the summer, when fish were moved along an
aluminum vessel, the fish swam the furthest into the electric barrier. During all seasons, fish that
were moved through the barrier along the canal wall advanced farther into the barrier before be-
coming incapacitated than those that were moved through the middle of the canal (Table 4). All
fish that were moved through the electric barrier recovered in less than one minute post-
incapacitation.
The PCA analysis revealed a temperature-conductivity gradient along the first principal
component, and revealed distinct groupings among fish sizes and the two environmental varia-
bles between seasons (Figure 33; Table 5). The second axis (PC2) showed a weak size gradient
(Figure 33; Table 5). The first and second principal component axes explained 72% and 24% of
the variation in the correlation matrix respectively. The fish used in the trial runs were largest in
the spring and winter when temperatures were lowest and conductivity was highest. The linear
regression between PC 1 scores and incapacitation distances revealed a significant inverse rela-
tionship (R2 = -0.60, p = 0.003; Fig. 34).
- 61 -
*P < 0.05
Table 4. Mean total length of gizzard shad used in trials, predicted distances at which 50% of the caged fish would be incapacitated, generalized coefficients of determination (R2) and estimated logistic regression slopes and intercepts under var-ious conditions. Results are only for trial runs that took place at the current barrier operating parameters.
- 62 -
Figure 32. Predicted relationships of the incapacitation probabilities of encaged Gizzard Shad as a function of distance during different seasons and alongside different boat hulls (spring and summer only [during the fall and winter, only aluminum-hull boats were used]). Lines extending from x and y axes de-note predicted point at which half of the fish became incapacitated. Points represent the means (± SE) of observations at site distances, but model is based on actual observations (0 or 1).
- 63 -
Figure 33. Principal components analysis bi-plot of three continuous variables measured during trial runs. Arrows represent eigenvectors multiplied by two to scale to bi-plot area.
- 64 -
Figure 34. Regressions of principal component 1 and barrier-run, incapacitation distances (natu-ral-log transformed). Diamonds (♦) denote trial runs along the canal wall with an aluminum-hull boat, squares (■) denote trial runs along the canal wall with a fiberglass-hull boat, triangles (▲) denote trial runs through the mid-channel of the canal with an aluminum-hull boat, and circles (●) denote trial runs through the mid-channel of the canal with a fiberglass boat.
Table 5. Correlation matrix of PC 1 and 2 scores with environmental and size variables.
Conductivity (mS/cm) Temperature (C) Fish Size (mm TL)Principal component 1 0.96* -0.89* 0.68*Principal component 2 -0.13 0.42 0.73*
Conductivity (mS/cm) -0.86* 0.54Temperature (C) -0.31
- 65 -
Six out of the eleven encaged Common Carp that were moved through the Demonstration
Barrier were not incapacitated. However, all of those Common Carp that were not incapacitated
did display reactions such as rapid circling motions within the cage and, in some cases, attempt-
ing to swim downstream as they were being moved through the barrier upstream. The two
Freshwater Drum that were moved through the barrier were not incapacitated nor did they dis-
play any type of behavior indicating that they were in distress (Table 6). A logistic regression
modeling the probability of incapacitation as a function of distance into the barrier was not sig-
nificant (slope P = 0.16). This is likely due to the small sample size entered into the model.
Table 6. Trial run information for large fish used in Demonstration Barrier caged-fish evaluations.
Run # Run location Species Length (mm TL) Incapacitated? Avoidance Behavior?1 Mid-channel barrier Common Carp 396 NO YES2 west wall Common Carp 490 NO YES3 Mid-channel barrier Freshwater Drum 371 NO NO4 west wall Common Carp 465 YES .5 Mid-channel barrier Freshwater Drum 384 NO NO6 west wall Common Carp 526 NO YES6 west wall Common Carp 518 NO YES7 Mid-channel barrier Common Carp 483 YES .7 Mid-channel barrier Common Carp 391 YES .7 Mid-channel barrier Common Carp 424 YES .8 west wall Common Carp 653 NO YES9 Mid-channel barrier Common Carp 589 NO YES9 Mid-channel barrier Common Carp 673 YES .
- 66 -
Discussion During the winter and spring, very few fish were present near the barriers and the ones
that were, were not present in the strongest parts of the barrier. During the summer, at both the
former and current operating parameters, more fish began to occupy the barrier area, yet there
was an abrupt decrease in the number of fish observed in the strongest parts of the barrier. How-
ever, both our standard and concentrated DIDSON sampling revealed that during the fall, the fish
were most abundant immediately below the barriers and more likely to actively challenge the
barriers. The large accumulations and challenging of the barriers occurred both at the reduced
and current operating parameters in the late summer/early fall. In fact, we observed no differ-
ences amongst individual fish distribution around the wide and narrow arrays in the fall (current
operating parameters). The distributions of fish schools during the fall, at the current operating
parameters, were not homogenous, but the differences between the sites within the zone of ulti-
mate field strength and the sites with low/no voltage was less distinct as in the other seasons.
This difference between the individual and schooled fish may be because the schooled fish tend-
ed to mostly occupy the canal walls, at the water surface, as shown with both the standard and
concentrated DIDSON sampling. Because the schools were concentrated in such small areas,
once they entered the zone of ultimate field strength, which were only sampled for ten minutes
during our standard DIDSON sampling, individual fish were observed to be more prevalent in
the zone of ultimate field strength than the schools. However, many more schools of fish within
the zone of ultimate field strength were observed than individuals. Holliman (2011) found that
schools of small Bighead Carp were more likely to challenge a laboratory barrier than individual
fish. Also, in a separate study, we observed numerous breaches of the CSSC barrier by schools
- 67 -
of fish, whereas individual, and even loosely aggregated fish were unable to breach the barrier
(Parker et al. 2013).
We found that although fish accumulated below the barrier in the summer, they were
homogenously distributed amongst site sections, indicating that they were most likely swimming
back and forth, probing the barrier, but not actively challenging it as much. However, in the fall,
fish were observed more near the canal walls, at the water surface. This is most likely because
the fish were more actively challenging the barrier, and that near the canal walls, at the surface,
is where the barrier voltage is weakest (Holliman et al. 2011). Also, the results of our caged-fish
work (discussed below) revealed that the fish always took slightly longer to become incapacitat-
ed when moved along the canal wall, compared to the mid-channel. Based on these results, the
barrier seems least susceptible to breach during the winter and spring. In the summer and fall
though, the barrier is particularly susceptible to breach because of the amount of fish present and
the increased activity levels. Increased activity and movement of temperate zone fishes in warm-
er weather is typical (e.g. Gauthreaux 1980). Because of this, it is recommended that any neces-
sary barrier maintenance shut-downs should be performed in the winter as much as possible, and
avoided during the summer and fall.
During our concentrated efforts, numerous instances of large schools of fish repeatedly
challenging the barrier were observed. We found that a combination of high water temperatures
coupled with dissolved oxygen concentrations at or greater than around 4 mg/l appears to explain
the high amount of barrier challenging behavior by resident fish of the CSSC. On several occa-
sions, fish were observed to be swimming through the entire DIDSON viewing cone that was
ensonifying the water surface within the zone of ultimate field strength. However, because a sin-
gle DIDSON viewing cone could not ensonify the entire width of the ultimate field strength
- 68 -
zone, no definitive conclusions could be made as to whether fish had actually breached the barri-
er by swimming all the way through it. However, recently, in the summer of 2013, two fixed-
position DIDSON units simultaneously recorded the entire width of the ultimate field strength
zone. With the two units, we recorded numerous instances of schools of small (<2in), wild fish
breaching the barrier (Parker et al. 2013).
Fish size was a significant factor affecting the degree to which wild fish could penetrate
into the barriers during the summer and fall with the smaller fish being able to penetrate the fur-
thest. During the winter and spring, we saw mostly larger fish, with the smallest ones being age-
1 fish from the previous year. In fact, as discussed in the materials and methods, the smallest
Gizzard Shad that we could locate in the spring were greater than 200 mm TL. However, during
the summer and fall is when that year’s young-of-the-year began occupying the canal and began
penetrating the barrier further than any other fish had. Holliman (2011) also found that smaller
Bighead Carp were able to penetrate further into an electrical barrier in a laboratory setting be-
fore movement was either inhibited or a breach occurred. The phenomenon of water-borne elec-
tricity affecting larger fish more than smaller fish has been well-established based on electrofish-
ing studies (Reynolds 1996; Ruetz et al. 2007).
The movement of caged fish alongside a boat for research purposes is still a rather novel
approach. The low number of fish that became incapacitated during our control trials validates
our methodology of moving fish through the barriers in cages without the cages having a signifi-
cant effect. Although all of the caged fish that we moved through the barriers were eventually
incapacitated, the distance into the barrier at which they were immobilized varied widely de-
pending on season and the relative conductivity of the hull type used.
- 69 -
After barrier operating parameters were increased, all fish were incapacitated in this
study. Variation in the distance at which the fish became incapacitated at the increased parame-
ters was dependent on a combination of water temperature, fish size, canal location (near wall or
mid-channel), and boat material the fish were next to during transport. Incapacitation occurred at
shorter distances into the electric barrier with larger fish and cooler temperatures (i.e., spring and
winter) relative to the smaller fish and warmer temperatures observed during the summer.
Conductivity in the CSSC is inversely related to temperature as a result of heavy winter
road salt application in the city of Chicago (Holliman 2011). Typically high conductivities de-
crease the effect of electricity on fish (e.g. Reynolds 1996). However, the cooler temperatures
likely compensate for the potential reduction in effectiveness, as evidenced by the shortest inca-
pacitation distances occurring during the highest conductivity observations.
We observed that the fish used in the winter were lethargic and swam very slowly during
the trials prior to incapacitation. We suggest that these results offer a line of evidence that the
barrier is least susceptible to breach during the winter. Reduced movement and activity of tem-
perate zone fishes during colder months has been well documented (e.g. Gauthreaux 1980). The
short distances into the barrier that the spring fish became incapacitated were likely influenced
by their larger size than cooler water temperature. Temperatures were higher in the spring rela-
tive to the fall, yet fish were incapacitated sooner in spring than fall. We used the smallest fish
that were available to us for the spring trials. However, the fish that we used were likely age-1
fish from the previous year. Our results indicate that larger fish size (>200 mm TL) is as effec-
tive at reducing fish penetration into the barrier as cold temperatures. Numerous studies relative
to the effects of electrofishing have revealed a positive relationship between fish size and electri-
cal effectiveness (e.g. Reynolds 1996).
- 70 -
The caged fish results during the summer showed the largest discrepancy relative to the
effect of the type of boat hull used for trials. Although the gizzard shad were of similar size, the
fish that were moved along the aluminum-hull boat were able to swim nearly twice the distance
into the barrier as those that were moved along the fiberglass-hull boat. Others have found that
steel-hulled barges and aluminum boats conduct electricity as they traverse the barrier in the
CSSC (Dettmers et al. 2005; McInerney et al. 2011; Slater et al. 2011), which can create voids of
non-electrified water that fish can swim unaffected in (Parker and Finney 2013). We also found
that, in all seasons, fish were able to swim into the barrier slightly farther along the canal wall,
indicating that the wall may be conducting electricity, and therefore weakening barrier effective-
ness as a fish deterrent or for immobilization.
The results of our caged-fish work at the Demonstration Barrier showed that caged fish
could pass without incapacitation. Dettmers et al. (2005) found that all caged fish > 150 mm TL
were incapacitated by the Demonstration Barrier along both the barge and a fiberglass boat.
However, when Dettmers et al. (2005) conducted their caged-fish trials in 2003, the Demonstra-
tion Barrier was operating at 0.39 V/cm. Today, as a result of age and deterioration, the Demon-
stration Barrier operates at about 0.31 V/cm (McInerney et al 2011). While six Common Carp
were moved through the barrier without becoming incapacitated, it is important to note that all of
those fish did exhibit “avoidance behavior,” such as rapid circling and attempting to swim down-
stream. The two Freshwater Drum, however, were not incapacitated, nor did they exhibit any sort
of “avoidance behavior” to indicate that they were in distress as they were being moved through
the Demonstration Barrier. Vulnerability to water-borne electricity can vary among species
(Henry et al. 2004) as a result of differences in morphology, physiology, and behavior (Reynolds
- 71 -
1996). Where Asian carp are in the gradient of electrical susceptibility, compared to Freshwater
Drum and Common Carp, is currently unknown.
The accumulation of fish below the barrier is not surprising given that fish migrate up-
stream in lotic systems for numerous reasons such as foraging, migration, reproduction, and es-
cape from predation (Northcote 1978; Dingle 1980; Stewart 1990). Our observations of the fish
continuously probing and challenging the barrier is consistent with other findings of fish repeat-
edly challenging an electrical barrier (Stewart 1981; Savino et al. 2001; Holliman 2011). In
some cases, fish that have a strong desire to move upstream will continuously challenge an elec-
trical barrier until they breach it, are harmed, or even killed (Stewart 1990; Bullen and Carlson
2003). Because of this, Stewart (1990) proposed that barriers should be used to modify a particu-
lar fish behavior rather than outright prevent it.
Both Bighead and Silver Carp have been shown to exhibit positive rheotaxis and make
long upstream dispersals (Peters et al. 2006; DeGrandchamp et al. 2008; Holliman 2011; Hoover
et al. 2012). Holliman (2011) found that when small Bighead Carp were placed in a laboratory
raceway with flowing water and a small electrical barrier that the fish continuously challenged
the barrier. Some fish even re-challenged the barrier immediately upon regaining muscle control
after being incapacitated and swept downstream out of the electrical field. This persistent chal-
lenging of the electrical field resulted in some fish breaching the barrier at the current operating
parameters of 0.91 V/cm (Holliman 2011).
The upstream extent of young-of-the-year Asian carp in the Illinois River, which would
pose the greatest threat to barrier breach, have been documented about 164 km below the barriers
(http://www.asiancarp.us/documents/ACDistribution.pdf). In December 2009 an adult Bighead
Carp was captured just below the barriers in the Lockport Pool during a rotenone event, and in
- 72 -
June 2010 an adult Bighead Carp was captured in Lake Calumet, which is 48 km upstream of the
electric barrier (Moy et al. 2011). Also, environmental DNA (eDNA) of both Bighead and Silver
Carp has been detected above the barrier system in the CAWS (ACRCC 2013; Jerde et al. 2011;
2013; Mahon et al. 2013).
The discovery of a small number of Asian carp further away from the larger, established
population is consistent with numerous other examples of leptokurtic dispersal patterns, in which
a few bold individuals move much further than the majority of the population (Skalski and
Gilliam 2000; Fraser et al. 2001; Rehage and Sih 2004; Roberts and Angermeir 2007) including
invasive fish (Jones and Stuart 2009; Lynch and Mensinger 2012). Furthermore, invasive species
often exhibit lag times from the period in which they are initially introduced to a novel habitat, to
when they disperse widely (Williamson 1996; Crooks and Soulé 1999; Nico and Fuller 1999)
and once established, recruitment may be variable (Madenjian et al. 2005; Weber and Brown
2013). This dispersal lag time phenomenon was displayed by Asian carp as well (Chick and Pegg
2001) and current recruitment in the Illinois River is variable (Irons et al. 2011). Asian carp ini-
tially invaded the lower Mississippi River in the 1970’s; however, they did not expand and in-
crease in large numbers in the Illinois River until the 1990’s (Chick and Pegg 2001).
Of great concern is when Barrier IIB (the upstream barrier) is de-energized, either as part
of a planned maintenance event or an unplanned outage. In order to properly maintain Barrier
IIB, maintenance must be performed at least four times per year. This requires keeping Barrier
IIA energized and then de-energizing Barrier IIB. Barrier IIA is downstream of site 2 in figs 2-7
and 2-8, meaning that any fish that are between those two barriers could opportunistically move
upstream. Others have found that fish will move upstream soon after an electrical barrier is de-
energized (Stewart 1981; Swink 1999; Godlewska et al. 2007; Holliman 2011). In the event that
- 73 -
fish opportunistically move upstream as a result of Barrier IIB being de-energized, movement
may still be inhibited by the Demonstration Barrier. However, the Demonstration Barrier is most
likely only effective on larger fish because of the age and subsequent deterioration of the struc-
ture (McInerney et al. 2011). Current management strategies to address this issue have been to
target and capture large fish (≥ 300 mm TL) that are between Barrier IIA and the Demonstration
Barrier following an outage or switch. However, these removal efforts have taken place any-
where from one to six weeks following an outage or barrier switch and have never removed all
of the large fish that were between the barriers.
Also of concern is the potential for barge vessels to facilitate fish passage beyond the bar-
riers. Besides the metal hull distorting the electric field (Dettmers et al. 2005; McInerney et al.
2011; Slater et al. 2011), barges also create a complex suite of hydrodynamic water motions as
they navigate through riverine waterways (Bhowmik and Mazumder 1990; Maynord and Siem-
sen 1990; Wolter and Arlinghaus 2003). The direct (Killgore et al. 2001; Gutreuter et al. 2003;
Wolter and Arlinghaus 2004; Killgore et al. 2011) and indirect (Gutreuter et al. 2006; Kucera-
Hirzinger et al. 2009) impacts of tow-barge vessel navigation on fish has been well investigated.
However, the actual distances that fish are physically displaced by barges, especially near elec-
trical barriers, is a topic that warrants further investigation. Currently, laboratory and field stud-
ies on the effects of barges on fish near the electric dispersal barrier are being conducted by
USACE and USFWS. Preliminary results have revealed multiple modes in which barge vessels
can facilitate fish breach (Parker and Finney 2013).
Briefly, as a barge moves through the water past a fixed point, it first creates a bow cur-
rent in front of the vessel followed by a bow wave. This bow wave creates a rise in the water
level in front of the vessel, which causes a water drawdown along the hull of the vessel. This
- 74 -
drawdown then creates a return velocity of water moving in the opposite direction of the vessel.
The water level then rebounds at the stern of the barge and the water directly behind the stern
travels in the direction of the barge vessel as wake flow. Finally, the water directly behind the
tow vessel is moved away from the tow by the propeller jet velocity (Bhowmik and Mazumder
1990; Maynord and Siemsen 1990; Wolter and Arlinghaus 2003). The end result of a tow-barge
vessel passing a fixed point is accelerated downstream flow if the vessel is navigating upstream,
or a temporary reverse flow if the vessel is navigating downstream (Bhowmik 1991). The magni-
tude and extent of these water movements is dependent on the size and displacement of the ves-
sel, relative to the water body, and the speed at which it is travelling (Wolter and Arlinghaus
2003). These barge-induced water movements are more exaggerated in a confined channel. A
confined channel is one which has a simple shape and has a significant cross-sectional area taken
up by navigating vessels (Martin 1997; Maynord 2004; Taylor et al. 2007) as opposed to a river-
ine system with sloped banks that can attenuate wave and flow energy. The section of the CSSC
that encompasses the electric barrier is rectangular in cross-sectional shape and is a confined
channel.
We observed small fish immediately downstream of the electrical field that encompasses
the ultimate field strength of 0.91 V/cm, which is about 2 meters wide at the surface. Larger fish
(>300 mm TL) were as close as 35 meters away from that field. Based on the complex barge hy-
drodynamics described above, both upstream and downstream-bound barges present different
modes of potentially moving a stunned fish past the electric barrier. Concerns about upstream-
bound barges are that the bow flow and bow wave may push fish into the electrical barrier to a
point at which they become incapacitated. Once incapacitated, the fish may then become en-
trained into one of the barge-induced flows. If a stunned fish were to remain directly in the path
- 75 -
of the barge until the barge hull physically struck it, the fish may remain entrained in the bounda-
ry layer (a layer of static water that moves in the same direction as the vessel [Martin 1997]) that
typically forms on the bottom of a barge hull. Also, if a stunned fish were to end up in the wake
flow behind the stern of the barge vessel it may remain entrained in that flow. Depending on the
amount of accelerated downstream flow though, some fish may be moved further downstream
away from the barrier, or stunned fish may end up getting moved back below the barrier after
initially being displaced into it. Parker and Finney (2013) found that fish that were struck by
barges as far as 100 m below the barriers were still entrained beyond the barriers.
Downstream-bound barges are also a concern because of the net reverse flow that they
create (Bhowmik 1991). This flow could potentially move fish into and beyond the barrier. We
have noted that Gizzard Shad used in the caged-fish trials quickly recover after becoming inca-
pacitated. Therefore, if a fish that is moved upstream of the barrier recovers while the flow is re-
versed, it may be able to continue swimming upstream. Of further concern is the low water ve-
locity that typically flows in the CSSC at the barrier site (~0.15-0.3 m/s). This low velocity can
take a long time to counteract the barge vessel-induced reverse flow. Also, the CSSC at the bar-
rier site often exhibits no flow or reverses on its own depending on how long the downstream
lock and dam structure is closed. We have observed reverse flows at the barrier site for as long as
one hour.
We found that once fish started penetrating upstream into the electric barrier, where the
electricity would be perceptible, that they were more prevalent along the canal walls, at the water
surface. The presence of these fish, which were actively challenging the barrier, at the water sur-
face is not surprising since the electrical field that is first emitted from the narrow barrier arrays
(on the bottom of the canal) is approximately 3.54 V/cm (Holliman 2011). We found that caged
- 76 -
fish took longer to become incapacitated when they were moved immediately adjacent to the ca-
nal wall through the barrier, indicating that the wall, may be conducting some of the electricity.
This may explain why the fish were more prevalent along the canal walls than in the middle of
the canal along the bottom. However, we did not assess fish abundance at the water surface in
the middle of the canal, therefore we do not know if fish were more prevalent near the walls at
the surface or if fish were homogenously distributed along the water surface. Although given the
caged-fish results, it seems likely that fish would prefer the wall, since it appears to weaken the
electrical field.
Small fish are easily displaced by navigation-induced waves (Holland 1986; Wolter and
Arlinghaus 2003; 2004; Kucera-Hirzinger et al. 2008). The effects of navigation-induced waves
are attenuated in embayments that are farther away from the main shipping channel or in more
complex habitat (Arlinghaus et al. 2002; Gabel et al. 2008). The vertical canal walls of the con-
fined CSSC are the simplest type of bankline that fish could occupy though. These simple
banklines cause the greatest amount of hydrologic momentum transfer (Myers 1978), which in-
creases the magnitude and duration of the displacement effects caused by barges. Specific navi-
gation-induced water motions that are more amplified in confined channels, compared to sloped
or protected shorelines, include increases in water drawdown, return velocity, transverse stern
waves, surge, and scour (Arlinghaus et al. 2002; Maynord 2004; Taylor et al. 2007). Because of
the increased magnitude, duration, and velocity of navigation-induced water movement around
vertical walls, particle displacement and sediment scour is typically greatest near the bankline
(Maynord 1990; Taylor et al. 2007). Therefore, even though the fish that we observed near the
canal walls were most likely not at risk of being directly struck by a barge vessel, they could still
- 77 -
be affected by the exaggerated water movements created along the walls. Of particular concern
are downstream-bound barge vessels which can create reverse flows (Bhowmik 1991).
Others have shown that fish are indeed entrained by towboats (Gutreuter et al. 2003; Mi-
randa and Killgore 2013), including Bighead and Silver Carp (Killgore et al. 2011). These stud-
ies have taken place in large rivers that are very different from the CSSC. Although Killgore et
al. (2011) found that at times, fish entrainment was high in narrow sections of water with slow
velocity. Indeed, the electrical electric barrier in the CSSC would present a different scenario in
which fish encounter a barge vessel than what others have studied. As we have shown, fish ac-
cumulate below the electrical barrier. The barrier may act as a “third wall” in the canal system
that does not allow fish to escape from oncoming barge vessels, thus making them more suscep-
tible to the barge-induced water movements. Clearly, much is still unknown about fish-barge in-
teractions at electrical barriers, which is why we feel that this is a topic worthy of future investi-
gations.
After the barrier operating parameters were increased from 0.79 to 0.91 V/cm, we found
that all fish used in our caged-fish trials were incapacitated, albeit at different distances into the
barrier depending on season, fish size, and the type of boat used to move them. Whereas, at the
previous parameters, eight fish were not incapacitated. Because, the operating parameters in-
creased soon after we began working at the electric barrier, we do not know how our results
would differ with the different settings over an entire year for both the caged-fish and the DID-
SON sampling. We were only able to sample the electric barrier during one season (summer)
when the operating parameters were at 0.79 V/cm, however, fish accumulation below the barriers
was similar at the same time the next year, at the increased parameters. In conclusion, we found
that wild fish accumulate below the electric barrier and at times will persistently challenge it.
- 78 -
Some of these fish, depending on their size, were able to penetrate into the barrier as far up as the
point of ultimate field strength. Initial results from current work, utilizing two DIDSON units
from a fixed platform, revealed numerous instances of schools of small fish breaching the barri-
er. Of particular concern is when Barrier IIB is de-energized and fish are allowed to move up-
stream, especially given the reduced operating capacity of the Demonstration Barrier. Also of
concern, and the subject of study at the time of this writing is the effects that barges may have on
displacing fish across the barriers. As previously stated, it is recommended that any necessary
barrier maintenance shut-downs should be performed in the winter as much as possible, and
avoided during the summer and fall.
Acknowledgements
This work was funded by the Great Lakes Restoration Initiative. We thank USACE –
Chicago and Rock Island Districts, the U.S. Coast Guard – Lake Michigan Sector, and the Met-
ropolitan Water Reclamation District of Greater Chicago for support. Numerous members of the
Columbia and LaCrosse FWS Fish and Wildlife Conservation Offices assisted in the field and
with DIDSON footage review. Brittany Ackerland, Nick Barkowski, Teresa Campbell, Jennifer
Johnson, Kenson Kanczuzewski, Jose Rivera, Matthew Shanks, Eric Stadig, and John West pro-
vided field support and reviewed DIDSON and caged-fish footage. Rikki Rogers produced the
schematics for Figures 2-2, 2-3, 2-5, and 2-6. Carl Ruetz III provided advice about statistical
analyses. Jarrell Smith calculated the vertical slice loss of the DIDSON cone for sections A and J
of our DIDSON sampling sites. Jan Dean and Carl Ruetz III provided valuable comments on a
separate journal manuscript about the caged-fish work, which were incorporated into this report.
- 79 -
References
Agresti, A. 2007. An introduction to categorical data analysis 2nd edition. John Wiley and Sons,
Hoboken, New Jersey.
Allison, P. D. 2008. Convergence failures in logistic regression. Proc. SAS Global Forum 360:1-
11.
Arlinghaus, R., C. Engelhardt, A. Sukhodolov, and C. Wolter. 2002. Fish recruitment in a canal
with intensive navigation: implications for ecosystem management. Journal of Fish
Biology 61:1386-1402.
Barwick, D. H. and L. E. Miller. 1996. Effectiveness of an electrical barrier in blocking fish
movement. Proceedings of the Annual Conference of the Southeastern Association of
Fish and Wildlife Agencies 50:139-147.
Becker, A. A. K. Whitfield, P. D. Cowley, J. Järnegren, and T. F. Næsje. 2011. An assessment of
the size structure, distribution and behavior of fish populations within a temporarily
closed estuary using dual frequency identification sonar (DIDSON). Journal of Fish
Biology 79:761-775.
Bhowmik, N. and B. S. Mazumder. 1990. Physical forces generated by barge-tow traffic within a
navigable waterway. Pages 604-609 in Chang, H. H. and J. C. Hill, editors. Hydraulic
engineering. American Society of Civil Engineers, New York.
Bhowmik, N. G. 1991. Hydraulic changes in rivers due to navigation. Pages 33-40 in Fan, S.-S.
and Kuo, Y.-H., editors. Proceedings from the 5th Federal Interagency Sedimentation
Conference.
Boswell, K. M., M. P. Wilson, and J. H. Cowan Jr. 2008. A semiautomated approach to
- 80 -
estimating fish size, abundance, and behavior from dual-frequency identification sonar
(DIDSON) data. North American Journal of Fisheries Management 28:799-807.
Breen, M. J. and C. R. Ruetz III. 2006. Gear bias in fyke netting: evaluating soak time,
fish density, and predators. North American Journal of Fisheries Management 26:32–41
Bullen, C. R. and T. J. Carlson. 2003. Non-physical fish barrier systems: their development and
potential applications to marine ranching. Reviews in Fish Biology and Fisheries 13:201-
212.
Burwen, D. L., S. J. Fleischman, and J. D. Miller. 2010. Accuracy and precision of salmon
length estimates taken from DIDSON sonar images. Transactions of the American
Fisheries Society 139:1306-1314.
Caine, L. P. 1978. Sanitation strategy for a lakefront metropolis: the case of Chicago. Northern
Illinois University Press, DeKalb, IL.
Chick, J. H. and M. A. Pegg. 2001. Invasive carp in the Mississippi River basin. Science 292:
2250-2251.
Clarkson, R. W. 2004. Effectiveness of electrical fish barriers associated with the Central
Arizona Project. North American Journal of Fisheries Management 24:94-105.
Cooke, S. L. and W. R. Hill. 2010. Can filter-feeding Asian carp invade the Laurentian Great
Lakes? A bioenergetics modeling exercise. Freshwater Biology 55:2138-2152.
Crooks, J. A. and M. E. Soulé. 1999. Lag times in population explosions of invasive species:
causes and implications. Pages 103-127 in Sandlund O. T., P. J. Schei and Å. Viken,
editors. Invasive species and biodiversity management. Kluwer Academic Publishers,
Dordrecht, The Neterlands.
- 81 -
Dawson, H. A., U. G. Reinhardt, and J. F. Savino. 2006. Use of electric or bubble barriers to
limit the movement Eurasian ruffe (Gymnocephalus cernuus). Journal of Great Lakes
Research 32:40-49.
DeGrandchamp, K. L., J. E. Garvey, and R. E. Colombo. 2008. Movement and habitat selection
by invasive Asian carps in a large river. Transactions of the American Fisheries Society
137:45-56.
Dettmers, J. M., B. A. Boisvert, T. Barkley, and R. E. Sparks. 2005. Potential impact of steel-
hulled barges on movement of fish across an electric barrier to prevent the entry of
invasive carp into Lake Michigan. Illinois State Natural History Survey, Center for
Aquatic Ecology, Aquatic Ecology Technical Report 2005/14, Urbana, IL.
Dingle H. 1980. Ecology and evolution of migration. Pages 2-101 in Gauthreaux Jr., S. A.,
editor. Animal migration, orientation, and navigation. Acedemic Press, London.
Fraser, D. F., J. F. Gilliam, M. J. Daley, A. N. Le, and G. T. Skalski. 2001. Explaining
leptokurtic movement distributions: intrapopulation variation in boldness and
exploration. The American Naturalist 158:124-135.
Gabel, F., X- F. Garcia, M. Brauns, A. Sukhodolov, M. Leszinski, and M. T. Pusch. 2008.
Resistance to ship-induced waves of benthic invertebrates in various littoral habitats.
Freshwater Biology 53:1567-1578.
Gauthreaux Jr., S.A. 1980. The influences of long-term and short-term climatic changes on the
dispersal and migration of organisms. In: Gauthreaux Jr., S.A. (Ed.) Animal migration,
orientation, and navigation. Academic Press, New York, New York, pp 104-159.
Godlewska, M., L. Doroszczyk, B. Długoszewski, and M. Mokwa. 2007. Acoustical monitoring
of fish at an electric barrier. Fisheries Bethesda.
- 82 -
Gotelli, N. J. and A. M. Ellison. 2004. A primer of ecological statistics. Sinauer Associates, Inc.,
Sunderland, Massachusetts.
Gutreuter, S., J. M. Dettmers, and D. H. Wahl. 2003. Estimating mortality rates of adult fish
from entrainment through the propellers of river towboats. Transactions of the American
Fisheries Society 132:646-661
Gutreuter, S., J. M. Vallazza, and B. C. Knights. 2006. Persistent disturbance by commercial
navigation alters the relative abundance of channel-dwelling fishes in a large river.
Canadian Journal of Fisheries and Aquatic Sciences 63:2418-2433.
Hadderingh, R. H. and H. Jansen. 1990. Electric fish screen experiments under laboratory and
field conditions. Pages 266-280 in Cowx I. G., editor, Developments in electric fishing.
Fishing News Books, Blackwell Scientific Publications, Oxford, England.
Hartley, W. G. and D. Simpson. 1967. Electric fish screens in the United Kingdom. Pages 183-
197 in R. Vibert editor. Fishing with Electricity – its applications to bioplogy and
management. Fishing News Books, Blackwell Scientific Publications, Oxford, England.
Henry, T. B., J. M. Grizzle, C. E. Johnston, and J. A. Osbourne. 2004. Susceptibility of ten fish
species to electroshock-induced mortality. Transactions of the American Fisheries
Society 133:649–654.
Holeck K. T., E. L. Mills, H. J. MacIsaac, M. R. Dochoda, R. I. Colautti, and A. Ricciardi. 2004.
Bridging troubled waters: Biological invasions, transoceanic shipping, and the Laurentian
Great Lakes. Bioscience 10:919-929.
Holland, L. E. 1986. Effects of barge traffic on distribution and survival of ichthyoplankton and
small fishes in the upper Mississippi River. Transactions of the American Fisheries
Society 115:162-165.
- 83 -
Holliman, F. M. 2011. Operational protocols for electric barriers on the Chicago Sanitary and
Ship Canal: influence of electrical characteristics, water conductivity, fish behavior, and
water velocity on risk for breach by small silver and bighead carp. March, 2011, Smith-
Root Inc, Vancouver, WA.
Holmes, J. A., G. M. W. Cronkite, H. J. Enzenhofer, and T. J. Mulligan. 2006. Accuracy and
precision of fish-count data from a “dual-frequency identification sonar” (DIDSON)
imaging system. ICES Journal of Marine Science 63:543-555.
Hoover, J. J., L. W. Southern, A. W. Katzenmeyer, and N. M. Hahn. 2012. Swimming
performance of Bighead Carp and Silver Carp: methodology, metrics, and management
applications. ANSRP Technical Notes Collection. ERDC/TN ANSRP-12-3. Vicksburg,
MS: U.S. Army Engineer Research and Development Center.
Irons, K. S., T. M. O’Hara, M. A. McClelland, and M. A. Pegg. 2002. White perch occurrence,
spread, and hybridization in the middle Illinois River, upper Mississippi River system.
Transactions of the Illinois State Academy of Science 95:207-214.
Irons, K. S., M. A. McClelland, and M. A. Pegg. 2006. Expansion of round goby in the Illinois
Waterway. The American Midland Naturalist 156:198-200.
Irons, K. S., G. G. Sass, M. A. McClelland, and J. D. Stafford. 2007. Reduced condition factor of
two native fish species coincident with invasion of non-native Asian carps in the Illinois
River, U.S.A. is this evidence for competition and reduced fitness? Journal of Fish
Biology 71 suppl D:258-273.
Irons, K. S. G. G. Sass, M. A. McClelland, and T. M. O’Hara. 2011. Bigheaded carp invasion of
the La Grange Reach of the Illinois River: insights from the long term resource
- 84 -
monitoring program. Pages 31-51 in D. C. Chapman and M. H. Hoff, editors. Invasive
Asian carps in North America. American Fisheries Society, Symposium 74, Bethesda,
MA.
Jerde, C. L., A. R. Mahon, W. L. Chadderton, and D. M. Lodge. 2011. “Sight-unseen” detection
of rare aquatic species using environmental DNA. Conservation Letters 4:150-157.
Jerde, C. L., W. L. Chadderton, A. R. Mahon, M. A. Renshaw, J. Corush, M. L. Budney, S.
Mysorekar, and D. M. Lodge. 2013. Detection of Asian carp DNA as part of a Great
Lakes basin-wide surveillance program. Canadian Journal of Fisheries and Aquatic
Sciences 70:522-526.
Jones, M. J. and I. G. Stuart. 2009. Lateral movement of Common Carp (Cyprinus carpio L.) in a
large lowland river and floodplain. Ecology of freshwater Fish 18:72-82.
Killgore, K. J., S. T. Maynord, M. D. Chan, and R. P. Morgan Jr. 2001. Evaluation of propeller-
induced mortality on early life stages of selected fish species. North American Journal of
Fisheries Management 21:947-955.
Killgore, K. J., L. E. Miranda, C. E. Murphy, D. M. Wolff, J. J. Hoover, T. M. Keevin, S. T.
Maynord, and M. A. Cornish. 2011. Fish entrainment rates through towboat propellers in
the upper Mississippi and Illinois Rivers. Transactions of the American Fisheries Society
140:570-581.
Kucera-Hirzinger, V., E. Schludermann, H. Zornig, A. Weissenbacher, M. Schabuss, and F.
Schiemer. 2009. Potential effects of navigation-induced wave wash on the early life
history stages of riverine fish. Aquatic Sciences 71:94-102.
Lynch, M. P. and A. F. Mensinger. 2012. Seasonal abundance and movement of the
invasive round goby (Neogobius melanostomus) on rocky substrate in the Duluth–
- 85 -
Superior Harbor of Lake Superior. Ecology of Freshwater Fish 21:64-74.
Maceina, M. J., J. W. Slipke, and J. M. Grizzle. 1999. Effectiveness of three barrier types for
confining grass carp in embayments of Lake Seminole, Georgia. North American Journal
of Fisheries Management 19:968-976.
Madenjian, C. P., T. O. Höök, E. S. Rutherford, D. M. Mason, T. E. Croley II, E. B. Szalai, and
J. R. Bence. 2005. Recruitment variability of alewives in Lake Michigan. Transactions of
the American Fisheries Society 134:218-230.
Mahon, A. R., C. L. Jerde, M. Galaska, J. L. Bergner, W. L. Chadderton, D. M. Lodge, M. E.
Hunter, and L. G. Nico. 2013. Validation of eDNA surveillance sensitivity for detection
of Asian carps in controlled and field experiments. PLoS One 8:e58316.
Martin, S. K . 1997. Physical model studies for riprap design of tow-induced forces. Report No.
WES/TR/CHL-97-7. Army Engineer Waterways Experiment Station, Vicksburg, MS.
Maynord, S. T. 1990. Velocities induced by commercial navigation. Technical Report HL-90-15,
Vicksburg, MS.
Maynord, S. T. 2004. Ship effects at the bankline of navigation channels. Maritime Engineering
157:93-100.
Maynord, S. T. and T. S. Siemsen. 1990. Physical forces generated by barge-tow traffic within a
navigable waterway. Pages 610-615 in Chang, H. H. and J. C. Hill, editors. Hydraulic
engineering. American Society of Civil Engineers, New York.
McInerney, M. K., B. S. Aubin, J. C. Trovillion, C. L. Baxter, E. T. Trovillion, V. F. Hock, Jr.,
and D. M. Weir. 2011. 2011 in-water testing of aquatic nuisance species dispersal
Barriers IIA and IIB with increased voltage and frequency operating parameters.
ERDC/CERL TR-11-23, Champaign, IL.
- 86 -
McLain, A. L. 1957. The control of the upstream movement of fish with pulsated direct current.
Transactions of the American Fisheries Society 86:269-284.
McMillan, F. O. 1928. Electric fish screen. Bureau of Fisheries Bulletin 44:97-128.
Mills, E. L., J. H. Leach, J. T. Carlton, and C. L. Secor. 1993. Exotic species in the Great Lakes:
A history of biotic crises and anthropogenic introductions. Journal of Great Lakes
Research 19:1-54.
Mills, H. B., W. C. Starrett, and F. C. Bellrose. 1966. Man’s effect on the fish and wildlife of the
Illinois River. Illinois Natural History Survey Biological Notes 57.
Miranda, L. E. and K. J. Killgore. 2013. Entrainment of shovelnose sturgeon by towboat
navigation in the Upper Mississippi River. Journal of Applied Ichthyology 29:316-322.
Moursund, R. A., T. J. Carlson, and R. D. Peters. 2003. A fisheries application of a dual-
frequency identification sonar acoustic camera. ICES Journal of Marine Science 60:678-
683.
Moy, P. B., I. Polls, and J. M. Dettmers. 2011. The Chicago Sanitary and Ship Canal aquatic
nuisance species dispersal barrier. Pages 121-137 in D. C. Chapman and M. H. Hoff,
editors. Invasive Asian carps in North America. American Fisheries Society, Symposium
74, Bethesda, MA.
Mueller, R. P., R. S. Brown, H. Hop, L. Moulton. 2006. Video and acoustic camera techniques
for studying fish under ice: a review and comparison. Reviews in Fish Biology and
Fisheries 16:213-226.
Myers, W. R. C. 1978. Momentum transfer in a compound channel. Journal of Hydraulic
Research 16:139-150.
Nico, L. G. and P. L. Fuller. 1999. Spatial and temporal patterns of nonindigenous fish
- 87 -
introductions in the United States. Fisheries 24(1):16-27.
Northcote, T. G. 1978. Migratory strategies and production in freshwater fishes. Pages 326-359
in S. D. Gerking, editor. Ecology of freshwater fish production. John Wiley and Sons,
Inc, New York, NY.
Parker, A.D. and S.T. Finney. 2013. Preliminary results of fish-barge interactions at the electric
dispersal barrier in the Chicago Sanitary and Ship Canal. U.S. Fish and Wildlife Service
Interim Report, 17 pg.
Parker, A.D., P.B. Rogers, S.T. Finney, and R.L. Simmonds Jr. 2013. Preliminary results of fixed
DIDSON evaluations at the electric dispersal barrier in the Chicago Sanitary and Ship
Canal. US Fish and Wildlife Service Interim report. 12 pages.
Peters, L. M., M. A. Pegg, and U. G. Reinhardt. 2006. Movements of adult radio-tagged Bighead
Carp in the Illinois River. Transactions of the American Fisheries Society 135:1205-
1212.
Rasmussen, J. L. 2002. The Cal-Sag and Chicago Sanitary and Ship Canal: A perspective on the
spread and control of selected aquatic nuisance species. U.S. Fish and Wildlife Service,
Report no. 4469. 26 pp.
Rasmussen, J. L., H. A. Reiger, R. E. Sparks, and W. W. Taylor. 2011. Dividing the waters: the
case for hydrologic separation of the North American Great Lakes and Mississippi River
Basins. Journal of Great Lakes Research 37:588-592.
Rehage, J. S. and A. Sih. 2004. Dispersal behavior, boldness, and the link to invasiveness: a
comparison of four Gambusia species. Biological Invasions 6:379-391.
- 88 -
Retzer, M. E. and B. Batten. 2005. Fishes of the Chicago region: A review of the Dennison and
Illinois Natural History Survey collections. Transactions of the Illinois State Academy of
Science 98:63-73.
Reynolds, J. B. 1996. Electrofishing. Pages 221-251 in B. R. Murphy and D. W. Willis, editors,
Fisheries Techniques, 2nd edition. American Fisheries Society, Bethesda, Maryland.
Roberts, J. H. and P.L. Angermeier. 2007. Spatiotemporal variability of stream habitat and
movement of three fish species. Oecologia 151:417-430.
Ruetz III, C. R., D. G. Uzarski, D. M. Krueger, and E. S. Rutherford. 2007. Sampling a littoral
fish assemblage: comparison of small-mesh fyke netting and boat electrofishing. North
American Journal of Fisheries Management 27:825–831
Sass, G. G. and B. C. Ruebush. 2010. An In-situ Test of the Aquatic Nuisance Species
Dispersal Barrier for Preventing Range Expansions of Small Fishes. Final Report on
Supplemental Research Contract USEPA-GLNPO EPA GL 9655501
Savino, J. F., D. J. Jude, and M. J. Kostich. 2001. Use of electrical barriers to deter movement of
round goby. Pages 171-182 in C. C. Coutant, editor. Behavioral technologies for fish
guidance. American Fisheries Society, Symposium 26, Bethesda, Maryland.
Schrank, S. J., C. S. Guy, and J. F. Fairchild. 2003. Competitive interactions between age-0
bighead carp and Paddlefish. Transactions of the American Fisheries Society 132:1222-
1228.
Skalski, G. T. and J. F. Gilliam. 2000. Modeling diffusive spread in a heterogeneous population:
a movement study with stream fish. Ecology 85:1685-1700.
Slater, M., N. Yankielun, J. Parker, and M. J. Lewandowski. 2011. CSSC Fish Barrier Simulated
Rescuer Touch Point Results, Operating Guidance, and Recommendations for Rescuer
- 89 -
Safety. Coast Guard Research and Development Center Report No. RDC-UDI-1142.
New London, CT.
Smith, J. M., Mather, M. E., Frank, H. J., Muth, R. M., Finn, J. T., McCormick, S. D. (2009).
Evaluation of a gastric radio tag insertion technique for anadromous river herring. N. Am.
J. Fish. Manage. 29:367-377.
Sparks, R. E., T. L. Barkley, S. M. Creque, J. M. Dettmers, and K. M. Stainbrook. 2010.
Evaluation of an electric fish dispersal barrier in the Chicago Sanitary and Ship Canal.
Pages 139-161 in D. C. Chapman and M. H. Hoff, editors. Invasive Asian carps in
North America. American Fisheries Society, Symposium 74, Bethesda, MA.
Steingraeber, M. T. and P. A. Thiel. 2000. The round goby (Neogobius melanostomus): another
unwelcome invader in the Mississippi River Basin. Transactions of the North American
Wildlife and Natural Resources Conference 65:328-344.
Stewart, P. A. M. 1981. An investigation into the reactions of fish to electrified barriers and
bubble curtains. Fisheries Research 1: 3-22.
Stewart, P. A. M. 1990. Electric screens and guides. Pages 140-156 in I.G. Cowx and P.
Lamarque, editors. Fishing with electricity, applications in freshwater fisheries
management. Fishing News Books, Oxford.
Stoeckel, J. A., R. E. Sparks, K. D. Blodgett, S. D. Whitney, and P. T. Raibley. 1996. Interbasin
dispersal of invading aquatic species. Illinois Natural History Survey Reports 424,
Champaign, IL.
Stoner, A. W. 2004. Effects of environmental variables on fish feeding ecology: implications for
the performance of baited fishing gear and stock assessment. Journal of Fish Biology
65:1445–1471
- 90 -
Swink, W. D. 1999. Effectiveness of an electrical barrier in blocking a sea lamprey spawning
migration on the Jordan River, Michigan. North American Journal of Fisheries
Management 19: 397-405.
Taft, E. P. 2000. Fish protection technologies: a status report. Environmental Science and
Policy 3:349-359.
Taylor, D., K. Hall, and N. MacDonald. 2007. Investigations into ship induced hydrodynamics
and scour in confined shipping channels. Journal of Coastal Research 50:491-496.
Verrill, D. D. and C. R. Berry Jr. 1995. Effectiveness of an electrical barrier and lake drawdown
for reducing common carp and bigmouth buffalo abundances. North American Journal of
Fisheries Management 15:137-141.
Weber, M. J. and M. L. Brown. 2013. Density-dependence and environmental conditions
regulate recruitment and first-year growth of common carp in shallow lakes. Transactions
of the American Fisheries Society 142:471–482.
Williamson, M. 1996. Biological Invasions. Chapman and Hall, London.
Wolter, C. and R. Arlinghaus. 2003. Navigation impacts on freshwater fish assemblages: the
ecological relevance of swimming performance. Reviews in Fish Biology and Fisheries
13:63-89.
Wolter, C. and R. Arlinghaus. 2004. A model of navigation-induced currents in inland
waterways and implications for juvenile fish displacement. Environmental Management
34:656-668.
- 91 -